Skip to main content

Changelog

All notable changes to the PostBoost API are documented here. Version numbers follow the OpenAPI spec and the SDK releases (the SDKs for a version are tagged v<version> in their repositories). The SDKs are released a few times a day at the latest API version, so a version released between two SDK releases has no SDK package of its own: the next SDK version includes its changes.

Base URL for all endpoints: https://postboost.co/app/api


v1.66.3 — 2026-10-23​

MCP​

  • The free tools server (/mcp/tools, postboost-tools) returns every tool in one tools/list response (up to 100 per page, was 50). Clients that read only the first page and ignore nextCursor now see all tools.

v1.66.2 — 2026-10-23​

What's New​

  • YouTube .sbv captions: captions downloaded from YouTube Studio as .sbv now lose their timing lines like .srt and .vtt files do, so the video tools read clean text instead of 0:00:00.599,0:00:04.160 lines.

API Changes​

  • generateSocialPostsFromVideo: the text description lists .sbv; SBV timing lines are removed before the length check and the AI call.
  • generateYoutubeTitles, generateYoutubeDescription: the transcript description lists .sbv; SBV timing lines are removed from the captions context.

MCP​

  • The generateSocialPostsFromVideo, generateYoutubeTitles and generateYoutubeDescription tools on /mcp/postboost (and their /mcp/tools twins) read .sbv captions the same way.

v1.66.1 — 2026-10-23​

What's New​

  • Agent skill builder: the API-key skill no longer tells agents to send an idempotency key the advanced MCP server does not accept. It now says to list the workspace's posts before retrying a create that may have gone through, and to treat a duplicate_post refusal as "the post already exists".

API Changes​

  • buildAgentSkill: for connection: api_token the idempotency guardrail line and the matching response example changed. The connector output is unchanged.

MCP​

  • buildAgentSkill returns the same new line for connection: api_token.

v1.66.0 — 2026-10-23​

What's New​

  • Captions as context for YouTube titles and descriptions: paste your video's captions (.srt, .vtt or the Show transcript text) next to the topic, and the titles and descriptions follow what the video actually says. Timestamps are removed and the first 6,000 characters are used; chapters still come only from your own chapter lines.

API Changes​

  • generateYoutubeTitles and generateYoutubeDescription: new optional request field transcript (up to 40,000 characters). Responses are unchanged.

MCP​

  • The generate-youtube-titles and generate-youtube-description tools take the same optional transcript.

v1.65.0 — 2026-10-22​

What's New​

  • Brand voice generator: a new free AI tool reads 3 to 10 of your own posts and returns your brand voice (personality, audience, topics, tone sliders, emoji and hashtag style, calls to action and example posts) with the quotes from your posts behind each trait, as a Markdown guide, an AI system prompt and JSON in PostBoost's brand voice format. Only quotes found in your posts are kept.

API Changes​

  • New generateBrandVoice (POST /tools/brand-voice-generator): send posts (3 to 10 different posts of 20 to 3,200 characters, 20,000 in total) and optional business_name and industry, used as given. It returns voice (the workspace brand voice fields, ready for PUT /brand-voice), evidence (up to 2 quotes per trait), system_prompt and markdown, plus credits_spent and cached. Nothing is saved. Needs a token and spends one AI text credit; the same posts and options again are free.

MCP​

  • New generateBrandVoice tool on /mcp/postboost, and generate-brand-voice on /mcp/tools (listed with a token).

v1.64.2 — 2026-10-22​

API Changes​

  • deleteMediaBulk and deleteMediaFolder (with delete_media: true) now say that deleting a media item also removes it from every post that uses it, including scheduled posts, which then publish without it or fail where the network requires media. Behaviour is unchanged.

MCP​

  • The deleteMediaBulk and deleteMediaFolder tool descriptions carry the same warning.

v1.64.1 — 2026-10-22​

API Changes​

  • previewPost: the X card now shows images whole from 3:4 to 2:1 (was 4:5 to 2:1), so min_ratio is 3:4 and a 3:4 photo is no longer reported as cropped. Taller images crop to 3:4.

v1.64.0 — 2026-10-22​

What's New​

  • YouTube video to social posts: a new free AI tool turns a video's captions (an .srt or .vtt file, or the text under Show transcript) into an X thread and up to three captions for LinkedIn, Instagram, Facebook, Threads, Bluesky or Mastodon in one run. Timestamps, cue numbers and sound tags are removed first. Each output fits its network's limit and links the video once.

API Changes​

  • New generateSocialPostsFromVideo (POST /tools/youtube-to-social-posts): send text (the captions) and optional video_url (a YouTube watch, youtu.be, Shorts or live link), title, outputs (1 to 4 of x_thread, linkedin, instagram, facebook, threads, bluesky, mastodon) and tone. It returns outputs (the thread as an array of posts, each caption as a string), outputs_skipped, warnings, the normalised video_url, cleaned_chars and removed_timestamps, plus credits_spent and cached. The video is never fetched. Needs a token and spends one AI text credit; the same captions and options again are free.

MCP​

  • New generateSocialPostsFromVideo tool on /mcp/postboost, and generate-social-posts-from-video on /mcp/tools (listed with a token).

v1.63.0 — 2026-10-22​

What's New​

  • LinkedIn carousel maker: a new free tool turns a post or an outline into LinkedIn carousel slides (1080×1350 or 1080×1080) and downloads them as one PDF for a document post, in the browser, with no watermark. Choose the most slides, a cover and a call-to-action slide, a light or dark theme, an accent colour and a footer with your name and @handle.

API Changes​

  • New buildLinkedinCarousel (POST /tools/linkedin-carousel-maker): send text and optional max_slides (3-20), size (4:5 or 1:1), cover and cta. It returns the slide outline: page (width and height in pixels), slides (n, kind cover/content/cta, heading, body lines), warnings (too_many_slides, split_block) and limits. The PDF itself is built on the tool page. Nothing is stored. Anonymous, within the standard free-tools limits.

MCP​

  • New buildLinkedinCarousel tool on /mcp/postboost, and build-linkedin-carousel on /mcp/tools.

v1.62.0 — 2026-10-21​

What's New​

  • Social media agent skill builder: a new free tool that builds a SKILL.md, AGENTS.md or system prompt for an agent that runs social media through PostBoost. You pick the tasks, the networks and how much the agent may do on its own (drafts only, ask before scheduling, or queue without asking); the file names the connector's real tools and adds guardrails such as a daily cap and validation before every write.

API Changes​

  • New buildAgentSkill (Free Tools API, POST /tools/social-media-agent-skill-builder): returns the file's content, filename, install steps, the connector tools it names (tools_used), its lines and warnings. No account needed.

MCP​

  • buildAgentSkill on /mcp/postboost and build-agent-skill on /mcp/tools return the same file.

v1.61.1 — 2026-10-21​

What's New​

  • Cross-post formatter, Threads links: links count once each, as Threads counts them (trailing punctuation trimmed). When a text splits into a Threads thread, each post is checked on its own, and the warning names the post that has more than 5 links.

API Changes​

  • formatCrossPost (Free Tools API): networks[].links is now the number of unique http(s):// links. On Threads, a text that splits into a thread gets one threads_too_many_links warning per post over 5 links, and none for the whole text.

MCP​

  • formatCrossPost on /mcp/postboost and format-cross-post on /mcp/tools return the same counts and warnings.

v1.61.0 — 2026-10-21​

What's New​

  • LinkedIn first-comment warning: when LinkedIn first comments are not enabled on the server, a LinkedIn post with a second content item now says so instead of dropping the comment silently. The post itself still publishes.

API Changes​

  • createPost, updatePost and validatePost: new warning code linkedin_first_comment_unavailable on the LinkedIn profile or Page account whose version has a first comment that will not be posted.
  • The content docs now state that LinkedIn posts the second item as the first comment only when first comments are enabled for LinkedIn on the server.

MCP​

  • createPost, updatePost and validatePost on /mcp/postboost return the same warning.

v1.60.0 — 2026-10-21​

What's New​

  • Threads link limit: Threads rejects a post with more than 5 different links. Creating, updating or validating a post now warns before Threads would reject it, and the pre-publish checker flags it on scheduled posts. Each post of a reply chain is checked on its own.

API Changes​

  • createPost, updatePost, validatePost: new warnings[].code value threads_too_many_links (a Threads account's post has more than 5 different http(s):// links). The post is still saved.
  • PostWarning.code is now an extensible list (x-extensible-enum instead of enum): new warning codes will be added without notice, so treat an unknown code as a warning and log it. In the generated SDKs code is a plain string, so a newer code no longer fails to deserialize in an older SDK.

MCP​

  • The validatePost tool on /mcp/postboost and the connector's post checks return the threads_too_many_links warning.

v1.59.0 — 2026-10-21​

What's New​

  • Giveaway comment picker: a new free tool draws giveaway winners and alternates from pasted comments or a CSV, in the browser, with no comment cap. It filters by one entry per person, keywords, a minimum number of @mentions and excluded usernames. The draw uses commit-reveal (pb-draw-v1): publish the seed's fingerprint when you announce the giveaway and the seed after the draw, and anyone can re-run it on the "Verify a draw" tab.

API Changes​

  • New pickGiveawayWinners (POST /tools/giveaway-comment-picker): send entrants (username, optional comment), an optional seed (64 hex characters), one_per_person, keywords, min_mentions, exclude, winners (1-100) and alternates (0-20). It returns winners, alternates, entrants_before, entrants_after, entrant_list (the canonical list to publish), entrants_hash, fingerprint, seed, pre_committed (false when the seed was generated for the call), algorithm and warnings (not_enough_entrants, not_pre_committed). Nothing is stored. Anonymous, within the standard free-tools limits and the 128 KB body cap.
  • countCharacters, splitThread and formatCrossPost: a text that starts with a byte-order mark (U+FEFF, which some copy and paste sources add) no longer fails on X. X counts ignore byte-order marks, and a link right after one is counted as plain text, as X's own counting rules do.
  • splitThread and formatCrossPost: non-breaking and ideographic spaces at the start or end of a paragraph are trimmed, and words split on them. The threads now match the ones the web tools show.

MCP​

  • New pickGiveawayWinners tool on /mcp/postboost, and pick-giveaway-winners on /mcp/tools.
  • The same fixes apply to the matching tools on /mcp/postboost and /mcp/tools.

v1.58.0 — 2026-10-21​

What's New​

  • Cross-post formatter: a new free tool shows one post on X, Threads, Bluesky, Mastodon and LinkedIn side by side. Each network counts with its own rules. The tool warns about more than 5 links on Threads, the LinkedIn "see more" fold and @handles that do not carry over, and gives a ready thread wherever the post is too long.

API Changes​

  • New formatCrossPost (POST /tools/cross-post-formatter): send text, optional networks (x, threads, bluesky, mastodon, linkedin; default all five, in the order you list them) and numbering. Each network returns count, limit, fits, over_by, links, warnings (threads_too_many_links, cut_by_fold, mentions_not_portable) and thread (the split posts when the text does not fit; null when it fits or on LinkedIn). Anonymous, within the standard free-tools limits.

MCP​

  • New formatCrossPost tool on /mcp/postboost, and format-cross-post on /mcp/tools.

v1.57.8 — 2026-10-20​

What's New​

  • Pixelfed expiry warnings: Pixelfed accounts now get the 7, 3 and 1-day expiry email and the account.token_expiring webhook before their access token ends (365 days after you connect by default; each server sets its own).

API Changes​

  • getAccountHealth: Pixelfed accounts report expiry_kind access with expires_at and days_left, instead of none.

v1.57.7 — 2026-10-20​

API Changes​

  • getXUsage: the description now says that scheduling to X is refused with 429 (code: x_monthly_cap_reached) only once enforcement is switched on (enforced: true). Until then usage is tracked only. No behaviour change.

v1.57.6 — 2026-10-20​

API Changes​

  • Social media post checker: the Instagram link_not_clickable warning now suggests the bio or a Story link sticker.

v1.57.5 — 2026-10-20​

API Changes​

  • countCharacters, checkPost and splitThread: the Mastodon count weighs only links that start with http:// or https:// as 23 characters. A bare domain such as example.com/blog/spring-launch now counts at its full length, as Mastodon counts it; before, it counted as 23, so a post near the limit could show as fitting and be refused by the server. X counting is unchanged.

MCP​

  • The connector tools count-characters and check-post, and the spec-generated tools for these operations, count Mastodon the same way.

v1.57.4 — 2026-10-19​

API Changes​

  • getSafeZones: the TikTok boxes are now measured from TikTok's own In-Feed safe-zone template, and source.basis says so: a 120 px action rail next to the 80 px right margin at 720 × 1280 (200 px from the edge), scaled to 1080 × 1920. The boxes are unchanged; last_verified is 2026-10-19.

v1.57.3 — 2026-10-19​

API Changes​

  • blogToSocial: captions now fit each network's real text limit, and character_limit reports it: Threads 500, Google Business Profile 1500, Bluesky 300, Pixelfed 500, Facebook groups 5000, and Instagram accounts connected with Instagram Login 2200, the same as other Instagram accounts. Before, Threads and Google Business Profile captions could run to 5000 characters and fail at publish.

v1.57.2 — 2026-10-19​

API Changes​

  • createPost, updatePost and validatePost: an Instagram account connected with Instagram Login (instagram_standalone) no longer gets a false instagram_image_ratio warning on a Story or Reel. The post type is now read from options.instagram_standalone.type for those accounts, as the publisher does. Their feed posts are still checked.

MCP​

  • createPost, updatePost and validatePost on /mcp/postboost return the corrected warnings.

v1.57.1 — 2026-10-19​

API Changes​

  • PostContent.url docs: the link also sets the Facebook Page link card (Facebook builds the card from the page's own Open Graph tags), not only LinkedIn and Bluesky.

v1.57.0 — 2026-10-18​

What's New​

  • Instagram bio, YouTube title and LinkedIn headline generators: three new free AI tools write five short drafts each, every one within the network's limit (150, 100 and 220 characters, counted like the character counter). The web pages also offer a formula and fill-in templates that work without AI.
  • YouTube description generator: a new free AI tool writes three description drafts with the hook in the first lines, your keywords used naturally and [link] placeholders instead of invented links. Each fits YouTube's limit of 5,000 UTF-8 bytes. Chapters are your own time and title lines, added as written; the tool never makes up times. The web page also offers a fill-in template that works without AI.
  • Social media manager pricing calculator on the Free Tools API: the maths behind postboost.co/tools/social-media-manager-pricing-calculator: monthly hours, a retainer, a per-post price and a ±20% range from your workload and your own hourly rate, with every assumption listed.
  • Webhook signature verifier: a new free tool checks an HMAC webhook signature (SHA-256, SHA-1 or SHA-512, hex or base64, optional prefix such as sha256=) against the raw body and secret. The web page computes it in the browser and sends nothing. When a signature does not match, it names the usual mistake that would make it match: a trailing newline, a re-encoded JSON body, a prefix left on, whitespace around the secret, the wrong encoding or algorithm. Presets for PostBoost (X-Signature) and GitHub (X-Hub-Signature-256).
  • Social media username generator on the Free Tools API: the generator behind postboost.co/tools/social-media-username-generator: 30 username ideas from a name or keywords, each within the chosen network's length and character rules, with a profile link per idea. Availability is not checked.
  • Social media glossary: more than 110 social media terms in plain language, from reach and engagement rate to ad metrics, webhooks and MCP, searchable by text and filterable by category. Each platform rule or number links to its source, and each term links to its place on the glossary page.
  • Alt text generator: a new free tool writes alt text for an image (short, up to 125 characters, or detailed, up to 250) plus a shorter alternative, so screen reader users get a description of what matters. The image is read once and never stored. The web page also has an alt text checklist that works without AI.
  • Social media report generator: a new free tool turns a CSV of post metrics into a one-page report: posts, views, engagements and engagement rate, a table per network, the top 5 posts and the best weekday. Copy it as text, download it as CSV or print it to PDF. The CSV is read in the browser and never uploaded; a CSV exported from PostBoost's reports maps with no changes.
  • Hashtag group builder: a new free tool cleans a messy list of hashtags or keywords (adds the missing #, cuts tags at punctuation, drops duplicates and number-only tags), checks the network's hashtag cap and splits the tags into copy-ready groups, on one line or as a first-comment block.
  • Social media tool picker: a new free tool matches what you need (a team dashboard, a REST API with an OpenAPI spec, SDKs, MCP, webhooks, a free hosted plan, a native Zapier, n8n or Make app) against PostBoost and the tools on our comparison pages, and shows what each one meets, misses or does not document, with the date the data was checked.
  • Social media automation recipes: eight step-by-step recipes for n8n, Zapier and Make that post through the PostBoost API with each platform's generic HTTP step: RSS to the queue, RSS to AI caption drafts, Google Sheets and Airtable rows to scheduled posts, Drive videos to Reels, TikTok and Shorts drafts, Slack approval, a failed-post alert and a weekly analytics digest. Each lists the API operations it calls; for n8n it also returns an importable workflow that uses only built-in nodes.
  • LinkedIn posts from the API: generateCaption now writes the same long-form LinkedIn posts as the LinkedIn post generator page. Pass network: "linkedin" and format: "post".

API Changes​

  • New generateBio (POST /tools/instagram-bio-generator): five Instagram bios by tone, with or without emoji and a call to action. Needs a token and spends one text credit per call, like generateCaption.
  • New generateYoutubeTitles (POST /tools/youtube-title-generator): five YouTube titles in one style (how-to, listicle, question, story, news), main keyword first, never in capitals and without < or >.
  • New generateLinkedInHeadline (POST /tools/linkedin-headline-generator): five LinkedIn profile headlines from a role, skills and an optional audience, by tone.
  • Each returns three to five variants; a reply with fewer usable ones is an error and spends no credit.
  • New generateYoutubeDescription (POST /tools/youtube-description-generator): three YouTube descriptions from a summary (up to 1,000 characters), optional keywords, optional chapter_lines (first at 0:00, at least three, in time order, at least 10 seconds apart) and a cta (subscribe, link or none). Needs a token and spends one text credit per call, like generateCaption. A reply with fewer than two usable descriptions is an error and spends no credit.
  • New calculateSmmPricing (POST /tools/social-media-manager-pricing-calculator): input networks, posts_per_week, content_type (copy, copy_design, copy_design_video), community_hours_per_week, reporting_hours_per_month, ad_hours_per_month, hourly_rate, currency (USD, EUR, GBP, formatting only) and pricing_model (hourly, retainer, per_post); output posts_per_month, hours_per_month, hourly_rate, monthly_retainer, per_post, range (low, high), line_items and assumptions. No token needed.
  • New verifyWebhookSignature (POST /tools/webhook-signature-verifier): returns valid, the expected signature and hints. Anonymous within the standard free-tool limits. It runs on our server, so send only test secrets and sample payloads; the body and secret are never logged or stored.
  • New generateUsernames (POST /tools/social-media-username-generator): input name (up to 60 characters), network (instagram, tiktok, x, youtube, threads), style (clean, creative, business, gaming), numbers and seed (the same input and seed return the same list); output usernames (username, profile_url) and the network's rules (min_length, max_length, separators). No token needed.
  • getGlossary (GET /tools/social-media-glossary): new free operation. Optional query q (text in the term or its definition) and category (metrics, content, ads, platforms, developer, strategy). No token needed; the standard free-tool limits apply.
  • New generateAltText (POST /tools/alt-text-generator): send the image as image_base64 in JSON, upload it as image in a multipart/form-data body, or give a public https image_url. JPEG, PNG, WebP or GIF (first frame) up to 5 MB; the type is read from the bytes, not the file name. Needs a token and spends one text credit per new image; the same image with the same options again is free.
  • This operation accepts request bodies up to 7 MB (the other free tools keep 128 KB). An image_url that points to a private address is refused, and each URL fetch counts against the daily limit of the tools that load a URL.
  • PostContent.url is now described: the link shown as a preview card on LinkedIn and Bluesky. When a content item has no url and no media, create, update and validate use the first link in body, as the web composer does, so API and RSS posts get the same link card. A url you send is never changed.
  • New summarizeReport (POST /tools/social-media-report-generator): send up to 5,000 rows, one object per post with date, network, text and the counts views, likes, comments, shares, clicks and engagements (whole numbers up to 100 billion; leave out the ones you do not have). The response has the totals, engagement_rate (null without views), networks, top_posts, weekdays and best_weekday. No token needed.
  • This operation accepts request bodies up to 2 MB (most free tools keep 128 KB). An invalid row is reported per field, for example rows.3.views.
  • New buildHashtagGroups (POST /tools/hashtag-group-builder): returns the clean hashtags, the removed entries with a reason (duplicate, digits_only, empty, too_long), the network's cap (null = no hard limit), the group_size used, the groups (tags, text, characters) and over_cap. Anonymous, within the standard free-tools limits.
  • New pickSocialMediaTool (POST /tools/social-media-tool-picker): send needs (or a use_case shortcut) and get every tool in results with its score, matched, missing and unknown needs, the compare_url of its full comparison, checked_at, pick_if and postboost_if, sorted by score, then name. Anonymous, within the standard free-tools limits.
  • findAutomationRecipes (GET /tools/social-media-automation-recipes): new free operation. Optional query platform (n8n, zapier, make), trigger (rss, spreadsheet, file, approval, webhook, schedule) and q (text in the title or goal). n8n_workflow is filled only when platform is n8n. No token needed; the standard free-tool limits apply.
  • generateCaption (POST /tools/instagram-caption-generator) takes an optional format: caption (default, unchanged) or post. A post is 300-600, 600-1,300 or 1,300-2,000 characters by length, with at most 3 hashtags. format: "post" works with network: "linkedin" only; any other network returns 422 and spends no credit.
  • A call without format gives the same captions as before.
  • getRemoteUploadStatus: the status values now match what the API sends: pending, downloading, processing, completed and failed (the spec said complete, which was never sent). The response documents progress (0 to 100) and error (when status is failed), and the new 404 response (status: failed with an error) for an unknown download id, one older than an hour, or one whose final result was already returned.
  • initiateRemoteUpload: the response is 200 (the spec said 201). Its body is status: completed with the media object, status: failed with an error, or status: pending with a download_id to poll. The spec described a bare Media object, which was never sent.
  • SDK upgrade: InitiateRemoteUpload201Response is now InitiateRemoteUpload200Response, and GetRemoteUploadStatus404Response is new. When polling getRemoteUploadStatus, check for completed (not complete).

MCP​

  • The three operations are new tools on /mcp/postboost and /mcp/tools.
  • The operation is a new tool on /mcp/postboost and /mcp/tools.
  • New calculateSmmPricing tool on /mcp/postboost, and calculate-smm-pricing on /mcp/tools.
  • New verifyWebhookSignature tool on /mcp/postboost, and verify-webhook-signature on /mcp/tools.
  • New generateUsernames tool on /mcp/postboost, and generate-usernames on /mcp/tools.
  • New getGlossary tool on /mcp/postboost, and get-glossary on /mcp/tools.
  • generate-alt-text is a new tool on /mcp/postboost and /mcp/tools. On /mcp/tools a request with a token may now be up to 1.5 MB, enough for an image of about 1 MB as image_base64; send larger images as image_url.
  • summarize-report is a new tool on /mcp/postboost and /mcp/tools. On /mcp/tools a request is capped at 128 KB without a token and 1.5 MB with one; send bigger reports to the API.
  • New buildHashtagGroups tool on /mcp/postboost, and build-hashtag-groups on /mcp/tools.
  • New pickSocialMediaTool tool on /mcp/postboost, and pick-social-media-tool on /mcp/tools.
  • New findAutomationRecipes tool on /mcp/postboost, and find-automation-recipes on /mcp/tools.
  • generate-caption on /mcp/tools and generateCaption on /mcp/postboost accept format.
  • The initiateRemoteUpload and getRemoteUploadStatus tools describe the same responses.

v1.56.0 — 2026-10-17​

What's New​

  • Social media holiday calendar on the Free Tools API: the data behind postboost.co/tools/social-media-holiday-calendar: public holidays, awareness days, weeks and months and social media days from November 2026 to December 2027, each with the official or organiser page its date comes from and a one-line post idea.

API Changes​

  • New listAwarenessDays (GET /tools/social-media-holiday-calendar): optional query from and to (YYYY-MM-DD; entries that overlap the range), country (a country code; worldwide entries are included), type (public-holiday, awareness-day, awareness-week, awareness-month, social-media-day) and topic; output covers (from, to), total, entries (name, type, date, end, countries, topics, source, note, idea), last_verified and sources. No token needed.

MCP​

  • New listAwarenessDays tool on /mcp/postboost, and list-awareness-days on /mcp/tools.

v1.55.1 — 2026-10-17​

What's New​

  • TikTok and LinkedIn presets for the AI writing tools: new free web tools (TikTok hashtag generator, TikTok caption generator, LinkedIn post generator) run the same generators as the Instagram tools. On the API, network: "tiktok" now tunes the output for TikTok the same way.

API Changes​

  • generateHashtags: with network: "tiktok" it returns 3 to 6 hashtags (was up to 10); other networks still get 5 to 10. The hashtags description says so.
  • generateCaption: with network: "tiktok" each caption opens with a short hook line. The captions description says so.

v1.55.0 — 2026-10-17​

What's New​

  • Social media audit scorecard on the Free Tools API: the logic behind postboost.co/tools/social-media-audit-template: answer an eight-question yes / partly / no checklist per network and get a 0-100 score per network and overall, the three weakest questions as top fixes with a tip each, and the scorecard as a Markdown report.

API Changes​

  • New scoreSocialAudit (POST /tools/social-media-audit-template): input target_posts_per_week (1 to 50, default 3) and one object per audited network (instagram, facebook, linkedin, x, tiktok, youtube, pinterest, threads, bluesky, mastodon) with yes, partly or no for profile, bio, link_tracking, frequency, replies, native_formats, pinned and engagement_rate (a missing answer counts as no); output score, networks (network, label, score, points, max_points, items with key, question, answer, points), top_fixes (key, question, tip, networks) and report_markdown. No token needed.

MCP​

  • New scoreSocialAudit tool on /mcp/postboost, and score-social-audit on /mcp/tools.

v1.54.0 — 2026-10-17​

What's New​

  • Dry run shows publish-time warnings: posts/validate now returns the same warnings as create-post, so a Threads post over 500 characters as Threads counts them, a YouTube description over 5,000 bytes, < or > in a YouTube title, or an Instagram feed image outside 4:5 to 1.91:1 shows up before you schedule instead of failing at publish.

API Changes​

  • validatePost: new required field warnings (array of PostWarning, always present, empty when there is nothing to say). valid keeps its meaning: warnings do not make a post invalid.

MCP​

  • The connector's validate-post tool returns warnings and its description says so.

v1.53.1 — 2026-10-17​

MCP​

  • The connector's get-post (with content_urls) returns url: null for a content item without a link, and get-account returns data: {} for an account without provider data, matching the REST API. Webhook payloads keep "" and [] for these two values; the webhooks page says how to read them.

v1.53.0 — 2026-10-17​

What's New​

  • Safe zone checker on the Free Tools API: the boxes behind postboost.co/tools/safe-zone-checker. Give a platform (TikTok, Instagram Reels or YouTube Shorts) and, optionally, a frame size, and get the areas the app's own interface usually covers plus a safe area for text and faces. The numbers are approximate: they come from each platform's guidance for vertical video ads, with the source and the date they were last checked in every answer.

API Changes​

  • New getSafeZones (POST /tools/safe-zone-checker): input platform (tiktok, instagram_reels or youtube_shorts) and optional width and height (default 1080 × 1920); output unsafe_areas (name, label, x, y, w, h), safe_area, source (label, url, basis), last_verified, caveat and warnings. Boxes are scaled from 1080 × 1920 and their edges rounded to the nearest pixel. An unknown platform returns 422. No token needed.

MCP​

  • New getSafeZones tool on /mcp/postboost, and get-safe-zones on /mcp/tools.

v1.52.0 — 2026-10-17​

What's New​

  • Influencer rate calculator on the Free Tools API: the logic behind postboost.co/tools/influencer-rate-calculator: give a network, follower count and deliverable, and get the sourced price-guide range for that tier next to the measured marketplace median, with deliverable, usage-rights and exclusivity adjustments and the source of every number.

API Changes​

  • New calculateInfluencerRate (POST /tools/influencer-rate-calculator): input network (instagram, tiktok, youtube, x, facebook), followers, optional deliverable, avg_views (required on YouTube), cpm_usd, usage_rights_months (0 to 24) and exclusivity; output tier, tier_label, deliverable, guide_low, guide_high (null when open-ended), guide_source, marketplace_median, marketplace_note, marketplace_source, cpm_estimate, multipliers, method, warnings, currency and last_verified. No token needed.

MCP​

  • New calculateInfluencerRate tool on /mcp/postboost, and calculate-influencer-rate on /mcp/tools.

v1.51.0 — 2026-10-17​

What's New​

  • Instagram carousel splitter on the Free Tools API: the cut-line maths behind postboost.co/tools/instagram-carousel-splitter. Give an image size, a slide count and a slide shape, and get the crop box of every slide in source pixels, so a wide image or panorama becomes a seamless carousel. Only the image size is sent, never the image.

API Changes​

  • New calculateCarouselSlices (POST /tools/instagram-carousel-splitter): input width and height (source pixels), slides (2 to 10) and ratio (1:1 or 4:5, default 4:5); output slides (x, y, width, height per slide, left to right), output (slide size in pixels), scale, upscaled and warnings. A source narrower than the slide count returns 422. No token needed.

MCP​

  • New calculateCarouselSlices tool on /mcp/postboost, and calculate-carousel-slices on /mcp/tools.

v1.50.0 — 2026-10-17​

What's New​

  • Posting frequency planner on the Free Tools API: the logic behind postboost.co/tools/posting-frequency-planner: give your networks and how many posts your team can make in a week, and get each network's sourced posts-per-week range (with source, date and evidence type), your posts split across the networks, a weekly grid and warnings.

API Changes​

  • New planPostingFrequency (POST /tools/posting-frequency-planner): input networks (1 to 9 of instagram, facebook, linkedin, x, tiktok, youtube, pinterest, threads, google_business_profile) and posts_per_week_capacity (1 to 200); output posts_per_week_capacity, sum_of_minimums, networks (network, label, recommended_per_week with min and max, null for an open-ended range, recommended_label, allocated_per_week, fit, source_type, sources), slots (weekday, network, label), warnings and last_verified. No token needed.

MCP​

  • New planPostingFrequency tool on /mcp/postboost, and plan-posting-frequency on /mcp/tools.

v1.49.0 — 2026-10-16​

What's New​

  • Post readability checker on the Free Tools API: the logic behind postboost.co/tools/post-readability-checker: give a post and get its Flesch reading ease and Flesch-Kincaid grade, the sentences of more than 25 words, possible passive phrases and plain suggestions. Hashtags, @mentions, links and emoji are not counted as words.

API Changes​

  • New checkReadability (POST /tools/post-readability-checker): input text (1 to 10,000 characters) and optional language (only en is scored; another language returns supported: false, a note and no scores); output supported, note, flesch_reading_ease, flesch_kincaid_grade, sentences, words, syllables, avg_words_per_sentence, long_sentences (index, words, excerpt), passive_hints, emoji_count, hashtag_count and suggestions. No token needed.

MCP​

  • New checkReadability tool on /mcp/postboost, and check-readability on /mcp/tools.

v1.48.0 — 2026-10-16​

What's New​

  • X API cost calculator on the Free Tools API: the logic behind postboost.co/tools/x-api-cost-calculator: give what your app posts and reads on X in a month and get the pay-per-use bill line by line, with the extra that link posts cost and the date X's prices were last checked.

API Changes​

  • New calculateXApiCost (POST /tools/x-api-cost-calculator): input posts, link_share (percent of posts with a URL, 0-100), replies, post_reads, user_reads, owned_reads (whole numbers, all optional); output line_items (count × unit price per resource), total_usd, link_premium_usd, notes, prices_last_verified and source. No token needed.

MCP​

  • New calculateXApiCost tool on /mcp/postboost, and calculate-x-api-cost on /mcp/tools.

v1.47.0 — 2026-10-16​

What's New​

  • Emoji search on the Free Tools API: the logic behind postboost.co/tools/social-media-emoji-picker: search 1,898 emoji (Emoji 15.1 and older, so they render on current phones) by Unicode name or CLDR English keyword, optionally within one of 9 categories.

API Changes​

  • New searchEmoji (POST /tools/social-media-emoji-picker): input query (up to 50 characters; empty lists the category in Unicode order), optional category and limit (1 to 100, default 30); output results (emoji, name, keywords, category, unicode_version) and total. An unknown query returns an empty results list, not an error. No token needed.

MCP​

  • New searchEmoji tool on /mcp/postboost, and search-emoji on /mcp/tools.

v1.46.0 — 2026-10-16​

What's New​

  • Social media ROI calculator on the Free Tools API: the logic behind postboost.co/tools/social-media-roi-calculator: give the revenue social media brought in and what it cost, and get the ROI, the net return and each cost's share, with the formula written out.

API Changes​

  • New calculateSocialRoi (POST /tools/social-media-roi-calculator): input revenue (USD), optional costs (ad_spend, tools, labor_hours, hourly_rate, content, other, all 0 or more) and period; output total_cost (labor = hours × hourly rate), net_return, roi_percent ((revenue − total cost) ÷ total cost × 100, 1 decimal, null when the total cost is 0), cost_breakdown, warnings and formula. No token needed.

MCP​

  • New calculateSocialRoi tool on /mcp/postboost, and calculate-social-roi on /mcp/tools.

v1.45.0 — 2026-10-16​

API Changes​

  • Posts (getPost, listPosts, createPost, updatePost and every other response that contains a post): a content item without a link returns url: null instead of an empty string, which the spec's uri format rejected. Requests still accept "".
  • Accounts (listAccounts, getAccount and accounts inside post responses, also for keys limited to some accounts): an account without provider data returns data: {} instead of [], matching the object type in the spec.
  • ScheduleResult.scheduled_at (schedulePost, addPostToQueue, approvePost) is documented in the format it has always had (Y-m-d H:i:s, UTC), not RFC 3339 date-time.
  • Workspace.created_at, User.created_at, User.email_verified_at, Subscription.trial_ends_at, paused_from, ends_at, Receipt.created_at and paid_at are documented as the display strings they have always been (for example Oct 16, 14:05 or Oct 16, 2026, in the configured timezone), not RFC 3339 date-time.
  • Upgrade note, typed SDKs: those 9 fields change type from a date to a string (Go time.Time → string, Java OffsetDateTime → String, Python datetime → str, TypeScript Date → string, PHP \DateTime → string). Go and Java could not parse these values before; TypeScript, Python and PHP code that treats them as dates must parse the string itself.
  • Upgrade note, untyped clients: a post content item's url can now be null. Code that runs string operations on it (an n8n expression, .length, .startsWith()) needs a null check.

v1.44.0 — 2026-10-16​

What's New​

  • Content calendar generator on the Free Tools API: the logic behind postboost.co/tools/social-media-content-calendar-generator: give a start date, the weeks to plan and posts per week per network, and get every post's date and time, spread evenly over each week, with your content pillars rotated over the posts, plus the plan as a CSV file and as an iCalendar (.ics) file.

API Changes​

  • New planContentCalendar (POST /tools/social-media-content-calendar-generator): input start_date (YYYY-MM-DD), weeks (1 to 12, default 4), posts_per_week (per network, 0 to 14), optional times (per network, HH:MM, default 10:00) and pillars (up to 6 themes); output end_date, total, slots (date, weekday, time, network, network_label, pillar), csv (UTF-8 with BOM) and ics (one 15-minute event per post, floating local time). A plan with no posts is a 422. No token needed.

MCP​

  • New planContentCalendar tool on /mcp/postboost, and plan-content-calendar on /mcp/tools.

v1.43.0 — 2026-10-16​

What's New​

  • Instagram image ratio warning: a post whose Instagram feed image is outside 4:5 to 1.91:1 (a 3:4 phone photo, a 9:16 screenshot) now gets a warning before it is scheduled, because Instagram rejects such images at publish time. The post is still created; crop the image to 4:5.

API Changes​

  • New warning code instagram_image_ratio in the warnings of post create and update responses: one per Instagram account whose feed post (single image or carousel) has images outside 4:5 to 1.91:1. Reels, Stories, videos and GIFs are not checked.

MCP​

  • The same warning in the create and update post tool results.

v1.42.1 — 2026-10-16​

API Changes​

  • countCharacters and checkPost now count Bluesky and Mastodon posts per grapheme, as the free tool pages do: every emoji counts 1, including Emoji 13 and later (🫶, 🥹, 🫠, 🩷, 🪿), skin-tone variants and new ZWJ sequences such as 🐦‍🔥, which counted 2 to 5 before. Links still count as before (Bluesky shortens them, Mastodon counts 23).

MCP​

  • The same counts on the countCharacters/checkPost tools of /mcp/postboost and count-characters/check-post on /mcp/tools.

v1.42.0 — 2026-10-16​

What's New​

  • Instagram line break generator on the Free Tools API: the logic behind postboost.co/tools/instagram-line-break-generator: give a caption and get it back with one spacer character on each blank line, so the gap between paragraphs stays when you paste it into Instagram, counted against the 2,200 character limit and the 5 hashtags PostBoost allows per Instagram post.

API Changes​

  • New formatInstagramCaption (POST /tools/instagram-line-break-generator): input text (up to 10,000 characters), optional spacer (braille = U+2800 Braille blank, default, or dot) and hashtags_to_end (move the hashtags that end a line to the bottom after three spacer lines); output text, count, limit, over, hashtags and max_hashtags. Spaces at line ends and blank lines at the start and end are removed. No token needed.

MCP​

  • New formatInstagramCaption tool on /mcp/postboost, and format-instagram-caption on /mcp/tools.

v1.41.0 — 2026-10-16​

What's New​

  • Blog to social on the Free Tools API: the AI behind postboost.co/tools/blog-to-social-media: give a blog post's URL (or its text) and up to 4 networks, and get one post per network, written with the same prompts as the web tool. Needs an API token; each post spends 1 AI credit of your workspace.

API Changes​

  • New generateSocialPostsFromBlog (POST /tools/blog-to-social-media): input url (public http or https) or text (up to 20,000 characters), platforms (1 to 4 of facebook, instagram, x, linkedin, pinterest, youtube, tiktok, mastodon), optional tone, length, hashtags, cta and workspace; output posts (platform, text, hashtags, characters, limit), source (title, final_url, only with url), warnings and credits_spent. Private, local and metadata addresses are a 422 with no request made. A call finishes within the request time budget: networks whose reply does not come back in time are left out with a platforms_skipped warning and are not charged; when none comes back the answer is a 502. Token only, like the other paid AI tools.
  • The workspace API's blogToSocial (drafts in a workspace) is unchanged.

MCP​

  • New generateSocialPostsFromBlog tool on /mcp/postboost, and generate-social-posts-from-blog on /mcp/tools (listed only with a token).

v1.40.0 — 2026-10-16​

What's New​

  • YouTube thumbnail checks on the Free Tools API: the rules behind postboost.co/tools/youtube-thumbnail-size-checker: give a thumbnail's width, height, file size and format and get a pass or fail per YouTube requirement. No account needed.
  • Social media video specs on the Free Tools API: the tables behind postboost.co/tools/social-media-video-specs: ratio, resolution, length, file size, formats and text safe zones for YouTube, Instagram, Facebook, TikTok, X, LinkedIn, Pinterest, Threads, Google Business Profile and Bluesky, with app and publishing API limits as separate rows and the sources of every row. No account needed.

API Changes​

  • New getVideoSpecs (GET /tools/social-media-video-specs): optional network (e.g. tiktok; an unknown key is a 422 that lists the allowed ones); output specs (network, placement, ratio, resolution, length, max_size, formats, safe_zone, basis, sources, and limits with the row as numbers: min_seconds, max_seconds, warn_over_seconds, min_ratio, max_ratio, recommended_ratio, max_mb), last_verified and sources. Standard Free Tools limits, no token.
  • New checkYoutubeThumbnail (POST /tools/youtube-thumbnail-size-checker): input width, height, file_size_bytes and format (MIME type), all required; output ok and checks (key = ratio, width, file_size or format; label; status = pass, fail or check; detail), against YouTube's 16:9, 640 px minimum width and 2 MB rules. Standard Free Tools limits, no token.

MCP​

  • New getVideoSpecs and checkYoutubeThumbnail tools on /mcp/postboost, and get-video-specs and check-youtube-thumbnail on the public /mcp/tools server.

v1.39.0 — 2026-10-16​

What's New​

  • Hook generator on the Free Tools API: the library behind postboost.co/tools/hook-generator: 83 hook templates for a post's first line, each with a style and the networks it suits, with your topic filled in. No account needed. With an API token, AI writes hooks for a specific topic, network and tone for 1 AI credit per call.

API Changes​

  • New getHookTemplates (GET /tools/hook-generator/library): optional topic (fills the {topic} placeholder), category (question, contrarian, list, story, mistake, how-to, curiosity, social-proof), network and search; output hooks (text, template, category, networks) and total. Standard Free Tools limits, no token.
  • New generateHooks (POST /tools/hook-generator): input topic, optional tone, count (1 to 10, default 5), network and workspace; output hooks (text, category), credits_spent and cached. Needs an API token and spends 1 AI credit of the workspace (0 for a cached answer). While AI is switched off it answers with count library templates and a notice instead of a 503, free of charge.

MCP​

  • New getHookTemplates and generateHooks tools on /mcp/postboost, and get-hook-templates and generate-hooks on the public /mcp/tools server (generate-hooks only with a token).

v1.38.0 — 2026-10-16​

What's New​

  • The spec now describes webhook deliveries: a WebhookEnvelope schema (event, id, occurred_at, workspace_uuid, data) and an OpenAPI 3.1 webhooks section with one entry per event, so you can type a webhook body from the spec instead of writing it by hand.

API Changes​

  • Breaking for SDK users (the HTTP API itself is unchanged):
    • Account.provider is a string: Java Account.getProvider() returns a String and the Account.ProviderEnum type is gone in Java and TypeScript. The old enum threw IllegalArgumentException (Java) on networks it did not list.
    • Post and CreatePost201Response: scheduled_at, published_at and created_at are strings (Y-m-d H:i:s, UTC) instead of date types such as Go time.Time or Java OffsetDateTime, which could not parse the values the API returns. Parse them with that layout if you need a date type. In Go the NewPost(...) constructor signature changes with them.
    • About 80 fields the API can return as null are now nullable in the SDKs, including MediaFolder.parent and AICredits.balance: Python Optional, TypeScript | null, PHP openAPINullables, and Go NullableX struct fields (the getters and setters keep working; direct struct field access does not).
  • The SDKs include the WebhookEnvelope model. The per-event webhooks entries stay in the spec and the docs but are left out of SDK generation, because they describe requests you receive, not calls you make. The webhook subscription methods (listWebhookSubscriptions, createWebhookSubscription, deleteWebhookSubscription) are unchanged.
  • Every field the API can return as null now uses the OpenAPI 3.1 form type: [X, 'null'] (87 fields that still used the 3.0 nullable keyword, which 3.1 validators and SDK generators ignore). Nullable enums list null as a value.
  • Account.provider is an open string with the known network keys in its description. The old enum listed 8 of the 17 keys, so strict SDK clients could fail to read an account list that holds, for example, a Bluesky, Threads or Google Business Profile account.
  • Post: scheduled_at, published_at and created_at are documented in the format they have always had (Y-m-d H:i:s, UTC), not RFC 3339 date-time.

MCP​

  • uploadReportBrandingLogo and downloadReportExport are no longer MCP tools: one needs a multipart file and the other returns a binary file. Use getReportExport, which returns a signed download URL.
  • Tool arguments the spec marks as nullable keep their real type: list arguments such as account_ids are arrays, numbers are numbers and booleans are booleans (some were offered as strings before).

v1.37.0 — 2026-10-16​

API Changes​

  • imageGenerate, imageEdit, imageVariations: the descriptions now state the real price, 1 image credit per image (429 image_credits_exhausted when the monthly pool is empty), with count × credit_weight text credits only while the image credit pool is switched off.
  • ImageGenerationResponse and ImageVariationsResponse now document image_credits_used and image_credits_remaining (nullable = no limit), which the endpoints already returned; credits_used is text credits only.
  • Media.id is documented as a string (a numeric id such as "42"): every media response has always returned it as a string, so typed SDK clients that expected an integer failed to read it.
  • Media: created_at is documented in the format it has always had (Y-m-d H:i:s, UTC), and thumb_url, width, height, duration_ms, alt_text use OpenAPI 3.1 null types so validators and SDKs accept the nulls the API returns.

v1.36.0 — 2026-10-16​

What's New​

  • Link preview checker on the Free Tools API: the checker behind postboost.co/tools/link-preview-checker loads a public page once and returns its Open Graph and X card tags, the title, description and image X, LinkedIn, Facebook and chat apps would show, and what to fix (missing or small image, long title or description, a canonical or redirect to another page, noindex). No account needed.

API Changes​

  • New checkLinkPreview (POST /tools/link-preview-checker): input url (http or https, up to 2,000 characters); output final_url, status, redirects, content_type, tags, image (declared or measured size), previews (x, linkedin, facebook, chat) and warnings. A private or non-http address, or a fourth redirect, returns 422; a page that does not load within 5 seconds returns 504. Results are cached for an hour. Limits: 20 calls per UTC day without a token, 200 with one.

MCP​

  • New checkLinkPreview tool on /mcp/postboost and check-link-preview on the public /mcp/tools server.

v1.35.0 — 2026-10-16​

What's New​

  • QR code generator on the Free Tools API: the generator behind postboost.co/tools/qr-code-generator returns a QR code as SVG for a link or text, with optional UTM tags added exactly as the UTM builder adds them, and warnings for low-contrast or inverted colours. No account needed.

API Changes​

  • New generateQrCode (POST /tools/qr-code-generator): input data (up to 1,000 characters), optional utm (source, medium, campaign, term, content), error_correction (L, M, Q, H), size (128-2048 px), foreground and background (#RRGGBB); output data (the encoded text), svg and warnings. A text too long for a QR code at the chosen level returns 422.

MCP​

  • New generateQrCode tool on /mcp/postboost and generate-qr-code on the public /mcp/tools server.

v1.34.0 — 2026-10-16​

What's New​

  • Threads counts emoji as UTF-8 bytes: Threads counts each emoji as its UTF-8 bytes toward the 500-character limit, so most emoji count as 4 and a family emoji as 18. Creating or updating a post now warns before Threads would reject it, the pre-publish checker flags it, and the free character counter, post checker and thread maker count Threads text the same way.

API Changes​

  • createPost, updatePost: new warnings[].code value threads_too_long (a Threads account's post is over 500 characters as Threads counts them). The post is still saved.
  • countCharacters, checkPost, splitThread: the threads row counts each emoji as its UTF-8 bytes.

MCP​

  • The countCharacters, checkPost and splitThread tools on /mcp/postboost, their /mcp/tools twins and the connector's character counter and post checker count Threads emoji as UTF-8 bytes.

v1.33.0 — 2026-10-16​

What's New​

  • Aspect ratio calculator on the Free Tools API: the maths behind postboost.co/tools/aspect-ratio-calculator: the simplified and decimal ratio of a width and height, or the missing side at a given ratio with the rounding stated, plus the social media image sizes with the same (or the nearest) ratio. No account needed.

API Changes​

  • New calculateAspectRatio (POST /tools/aspect-ratio-calculator): input width + height, or ratio (e.g. 16:9, 1.91:1) with one of them; output ratio, decimal, width, height, rounding, orientation and matches (presets from getImageSizes).

MCP​

  • New calculateAspectRatio tool on /mcp/postboost and calculate-aspect-ratio on the public /mcp/tools server.

v1.32.0 — 2026-10-16​

What's New​

  • Social media post preview on the Free Tools API: the fold and image data behind postboost.co/tools/social-media-post-preview: for Instagram, X, LinkedIn and Facebook, the caption count against the limit, the text the feed shows before its "more" link and the image ratios the feed shows uncropped. Give the image size to learn the ratio each card shows and whether it crops. Fold positions and X's image range are approximate. No account needed.

API Changes​

  • New previewPost (POST /tools/social-media-post-preview): input text, optional instagram_ratio (1:1, 4:5 or 1.91:1) and optional image_width + image_height (pixels, both or neither); output networks, one entry per card with key, count, limit, over, preview, folded, min_ratio and max_ratio, plus display_ratio and cropped when the image size is given.

MCP​

  • New previewPost tool on /mcp/postboost and preview-post on the public /mcp/tools server.

v1.31.0 — 2026-10-15​

What's New​

  • LinkedIn and YouTube connections are checked daily too: the dead-token check now covers LinkedIn profiles and Pages (token introspection) and YouTube (Google tokeninfo), so a revoked connection shows "needs reconnect" before the next post fails.

API Changes​

  • getAccountHealth: each account's connection has a new reconnect_reason (code, message, found_at) when the daily token check found its token dead, else null. Codes: meta_190_<subcode>, refresh_expired, linkedin_revoked, linkedin_expired, google_revoked.

MCP​

  • getAccountHealth on /mcp/postboost and get-account-health on the connector return connection.reconnect_reason.

v1.30.0 — 2026-10-15​

What's New​

  • Image sizes, crop maths and best times to post on the Free Tools API: the data behind postboost.co/tools/social-media-image-sizes and /tools/best-time-to-post, with its sources and last-verified date, plus the resizer's crop calculation. No account needed.

API Changes​

  • New getImageSizes (GET /tools/social-media-image-sizes?network=): every size row with network_key, format_key, pixel width/height when fixed, ratio, limits, basis and source, plus last_verified and sources.
  • New calculateCrop (POST /tools/social-media-image-sizes/crop): input width, height, then network + format (keys from getImageSizes) or target_width + target_height, and fit (cover or contain); output crop, output, image, scale and upscaled.
  • New getBestTimesToPost (GET /tools/best-time-to-post?network=): one table per network (a column per study, a row per weekday), last_verified, sources and a pointer to getBestPostingTimes for times from your own accounts.
  • The two GET operations send Cache-Control: public, max-age=86400 and an ETag; If-None-Match with it returns 304. An unknown network or format is a 422 that lists the allowed values.
  • listFreeTools lists one entry per operation: a tool with several operations appears once per operation_id, each with its own docs_url.

MCP​

  • New get-image-sizes, calculate-crop and get-best-times-to-post tools on the public /mcp/tools server and on /mcp/postboost.

v1.29.1 — 2026-10-15​

What's New​

  • AI free tools match the web pages: the hashtag and caption generators accept the same input as the tools on postboost.co, and a generated caption always fits the chosen network's text limit (at most 2,200 characters; 280 for X).

API Changes​

  • generateHashtags: topic is at most 200 characters (was 1,000); niche is one of the page's niches (Fitness, Food and restaurants, ...; was free text).
  • generateCaption: text is at most 500 characters (was 1,000); each of the three captions fits the network's text limit, cut at the last sentence end that fits.
  • Both: 503 now means AI tools are switched off; a call with a token is no longer refused when the guest AI budget is spent (your AI credits pay for it, as on the web pages).

MCP​

  • generateHashtags / generate-hashtags and generateCaption / generate-caption on /mcp/postboost and /mcp/tools take the same limits.

v1.29.0 — 2026-10-15​

What's New​

  • AI free tools on the API: the hashtag and caption generators from postboost.co/tools now run from code for any of Instagram, TikTok, LinkedIn, X, Threads and Facebook. They need an API token; a new generation spends 1 AI credit of your workspace, and repeating the same request is free (cached).

API Changes​

  • New POST /tools/instagram-hashtag-generator (generateHashtags): topic, optional niche, network, workspace; returns hashtags, credits_spent, cached.
  • New POST /tools/instagram-caption-generator (generateCaption): text, tone, length, emoji, hashtags, network, workspace; returns three captions, credits_spent, cached.
  • Both: 401 without a token (body has register_url), 403 for a workspace you do not belong to or for a scoped API key without the ai ability or bound to another workspace (code: insufficient_scope; a scoped key pays with its own workspace), 429 with code: ai_credits_exhausted and upgrade_url when the workspace has no AI credits left, 502 when the AI provider fails (no credit spent), 503 when AI tools are off or out of budget.
  • FreeToolError has a new optional code field.

MCP​

  • New tools generateHashtags and generateCaption on /mcp/postboost, and generate-hashtags and generate-caption on /mcp/tools (listed there only with a valid token). Scoped API keys get the same insufficient_scope checks as on the API.

v1.28.0 — 2026-10-15​

What's New​

  • YouTube description limit in bytes: YouTube accepts at most 5,000 bytes in a video description, and emoji or non-Latin letters take 2 to 4 bytes each. Creating or updating a post now warns before the upload would fail, and the free character counter counts the YouTube description in bytes.

API Changes​

  • createPost, updatePost: two new warnings[].code values: youtube_description_too_long (the description of a YouTube account's version is over 5,000 UTF-8 bytes) and youtube_forbidden_characters (the YouTube title or description contains < or >). The post is still saved.
  • countCharacters: the youtube row (video description) counts UTF-8 bytes; youtube_title still counts characters.

MCP​

  • The countCharacters tool on /mcp/postboost and count-characters on /mcp/tools and the connector count the YouTube description in bytes.

v1.27.1 — 2026-10-15​

What's New​

  • Free Tools API docs: one page per free tool with curl and 9 SDK examples, a "Free tools in 60 seconds" quickstart (no token, then with a token) and the public MCP server https://postboost.co/mcp/tools in the MCP guide. Free-tool calls work without a token in every SDK; to use a token, send it as a default Authorization header (the quickstart shows how per SDK).

API Changes​

  • listFreeTools: docs_url now links each tool's docs page (it was null).

v1.27.0 — 2026-10-15​

What's New​

  • Social media post checker: POST /app/api/tools/social-media-post-checker (checkPost) checks a post for each network before you publish: length as the network counts it, hashtag limit, media the network requires (Instagram, TikTok, Pinterest, YouTube), media count and type, video length, the see-more cut-off and links that are not clickable. Errors block publishing, warnings do not. No account needed. Same logic as the free tool at https://postboost.co/tools/social-media-post-checker.

API Changes​

  • New free tool operation checkPost with FreeToolCheckPostRequest / FreeToolCheckPostResponse; its input has an array (networks) and an object (media).

MCP​

  • /mcp/postboost: new tool checkPost (in-process).
  • /mcp/tools (public, no token): new tool check-post.
  • Connector (/mcp): new read-only tool check-post.

v1.26.2 — 2026-10-15​

API Changes​

  • Free Tools API (countCharacters, splitThread, cleanText and the other POST tool operations): request bodies up to 128 KB are accepted (was 64 KB), so the documented 10,000-character text fits even when your JSON encoder escapes every emoji to a 12-byte 😀 pair (Python json.dumps, PHP json_encode defaults). Larger bodies are still a 413.

MCP​

  • /mcp/tools: the same 128 KB body limit (was 64 KB).

v1.26.1 — 2026-10-15​

API Changes​

  • createConnectLink: the description's first line now carries the whole instruction (open the link, wait for the user, then list-accounts).
  • The API overview and the blogToSocial custom_instructions example no longer state a fixed platform count.

MCP​

  • /mcp/postboost: the createConnectLink tool description is no longer cut off mid-sentence.
  • The connector (POST /mcp) builds links and URLs from the configured site URL instead of the request's Host header.

v1.26.0 — 2026-10-15​

What's New​

  • AI text cleaner: POST /app/api/tools/ai-text-cleaner (cleanText) replaces em dashes, curly quotes, the ellipsis character, invisible characters and markdown markers in AI-written text, and flags common AI phrases with their offsets without changing them. No account needed. Same logic as the free tool at https://postboost.co/tools/ai-text-cleaner.

API Changes​

  • New free tool operation cleanText with FreeToolCleanTextRequest / FreeToolCleanTextResponse.

MCP​

  • /mcp/postboost: new tool cleanText (runs in-process, no HTTP sub-request).
  • /mcp/tools (public, no token): new tool clean-text.

v1.25.0 — 2026-10-15​

What's New​

  • Free tools over MCP, no account needed: a public MCP server at https://postboost.co/mcp/tools lists every free tool with an API handler (character counter, X thread splitter, LinkedIn formatter, fancy text, UTM builder, engagement rate). Calls without a token have per-IP limits; a limit error includes a register_url.

MCP​

  • /mcp/tools (new, no auth): one tool per free tool, named after its operationId in kebab-case (count-characters, split-thread, ...). A bearer token is optional; an invalid one is a 401. Bodies over 64 KB are a 413.
  • /mcp/postboost: the free tool operations (countCharacters, splitThread, ...) now run in-process instead of calling the REST API over HTTP, with the same limits and usage counting.
  • Connector (/mcp): new read-only tool count-characters.

v1.24.1 — 2026-10-15​

API Changes​

  • countCharacters: without an API token, text takes up to 5,000 characters (longer text is a 422); with a token the limit stays 10,000. Counting all networks is now about twice as fast.

MCP​

  • The count-characters tool description names the 5,000-character limit for calls without a token.

v1.24.0 — 2026-10-15​

What's New​

  • Connect new accounts from an agent: an agent asks for a single-use link, the user opens it while logged in to PostBoost, confirms, and the usual sign-in with the network connects the account. The link expires after 15 minutes and signs no one in.

API Changes​

  • createConnectLink: new POST /{workspaceUuid}/accounts/connect-links (body provider), returns 201 with url, expires_at and provider; 422 for Telegram, Discord and Slack (use connectChannel) and for Mastodon, Pixelfed and Bluesky (connected in the app). Workspace admins only, 10 links per minute.

MCP​

  • New create-connect-link tool on the connector (/mcp) and createConnectLink on the advanced server.

v1.23.0 — 2026-10-14​

What's New​

  • Free Tools API limits: calls without a token get 30 per minute and 500 per day per IP address; API tokens get 300 per minute, and Free-plan users 1,000 calls per day (paid plans have no daily cap).

API Changes​

  • Every Free Tools operation (listFreeTools, countCharacters, splitThread, formatLinkedInPost, generateFancyText, buildUtmUrl, calculateEngagementRate): responses carry X-RateLimit-Limit and X-RateLimit-Remaining; 429 adds Retry-After and the body fields retry_after, register_url or upgrade_url.

v1.22.0 — 2026-10-14​

What's New​

  • Free Tools API: the free tools at postboost.co/tools are now API calls under /tools: character counter, thread maker, LinkedIn formatter, fancy text, UTM builder and engagement rate. No account needed: call them without a token, or with any API token (no workspace, plan or api capability). Same answers as the web tools. Calls are limited to 30 per minute per IP for now, with or without a token.

API Changes​

  • New tag Free Tools, generated from the tool handlers by php artisan tools:spec: listFreeTools (GET /tools), countCharacters, splitThread, formatLinkedInPost, generateFancyText, buildUtmUrl and calculateEngagementRate (POST /tools/{slug}), each marked x-free-tool: true, with FreeTool*Request / FreeTool*Response schemas and examples.
  • These operations accept requests without a token (security: []); a token is optional, and an invalid token is 401. Bodies over 64 KB are 413, invalid input is 422 (countCharacters and splitThread take up to 10,000 characters), too many requests are 429.

MCP​

  • The advanced server (/mcp/postboost) gains one tool per new operation.

v1.21.0 — 2026-10-14​

What's New​

  • Failed-post fixer agent: a new agent type, failed_post_fixer, looks at posts that failed on an account in the last lookback_hours and proposes the next step: reconnect_account when the account must be reconnected, otherwise retry_account, with the caption cut to the network limit (counted the way the network counts) when the text is over it, found by PostBoost's pre-publish check or the network's error. One open proposal per post and account: a newer one replaces the older (expired). No AI, no credits; every retry waits for a person's approval.

API Changes​

  • WorkspaceAgent.type and WorkspaceAgentInput.type: new value failed_post_fixer. WorkspaceAgentSettings: new lookback_hours (1 to 72, default 24).
  • AgentProposal.kind: new values retry_account (payload error and, for a shortened caption, source_account_id, item_index, before, after) and reconnect_account (payload error, reconnect_url).
  • approveAgentProposal: retry_account saves the shortened caption on that account's version and publishes the account again; it becomes stale when the account is no longer failed, the post is publishing or the plan does not allow scheduling. reconnect_account only records the decision.

MCP​

  • The advanced server's create-agent and update-agent accept failed_post_fixer and lookback_hours. The connector tools list-agent-proposals and approve-agent-proposal describe the two new kinds.

v1.20.0 — 2026-10-14​

What's New​

  • Warnings on post create and update: a successful createPost or updatePost can now say when something will not behave as you expect, without failing the call: a warnings array of {code, account_id, message}. The key is absent when there is nothing to report, so existing clients see no change. See Guardrails → Warnings.

API Changes​

  • New schemas PostWarning and PostWarnings. createPost 201 and updatePost 200 may include warnings. updatePost 200 now documents its own body (success plus warnings).
  • Codes: ai_media_without_disclosure (AI-generated media with ai_generated: false, so no AI label) and meta_first_comment_unavailable (a Facebook Page or Instagram first comment while Meta reviews our access).

v1.19.0 — 2026-10-14​

What's New​

  • Long repurpose runs no longer time out: repurposeContent (POST /{workspaceUuid}/ai/repurpose) queues the run and waits up to 20 seconds. A run that finishes in time answers 200 with the same body as before. A slower run answers 202 with a run_id and status_url instead of a gateway error. Poll the new getRepurposeRun until the status is ok, partial or failed.

API Changes​

  • repurposeContent: new optional body field async (boolean, default false) to get the 202 right away; new 202 (RepurposeRunAccepted) and 500 (the run failed while the call waited, with its run_id) responses. Validation, account and credit errors still answer 422/429 before anything is queued.
  • New getRepurposeRun (GET /{workspaceUuid}/ai/repurpose/runs/{runId}, tag AI, 60 calls per minute): status, error, post_uuid, credits_used and, for one hour after a successful run, result (the 200 data body). A run still queued or running after 10 minutes reports failed.

MCP​

  • The MCP connector's repurpose-content tool tells you to call the new read-only get-repurpose-run tool with the run_id when a run takes longer than the wait. The advanced server gains get-repurpose-run from the spec.

v1.18.0 — 2026-10-14​

What's New​

  • RSS feeds can wait for approval: mode: needs_approval on createRssFeed/updateRssFeed gives each new item's post the next free queue slot with status "needs approval"; nothing publishes until someone approves it. See Automations.

API Changes​

  • mode on RSS feeds (createRssFeed, updateRssFeed, RssFeed) accepts needs_approval. With a scoped API key it needs write, like draft.

MCP​

  • Connector add-rss-feed lists needs_approval in its mode enum.

v1.17.3 — 2026-10-14​

What's New​

  • Free plan daily caps: Free workspaces get half the per-account daily caps (rounded down, at least 1), for example 12 Instagram posts per UTC day instead of 25. Paid plans are unchanged. See Guardrails.

API Changes​

  • No schema changes. account_daily_cap_reached errors and daily_cap.limit in GET /{workspaceUuid}/account-health and validatePost show the reduced cap.

MCP​

  • The post tools and the Claude connector refuse scheduling past the reduced cap the same way.

v1.17.2 — 2026-10-14​

What's New​

  • Clear errors for malformed schedule input: schedulePost and approvePost answer 422 with validation errors instead of 500 for a malformed date, time, timezone or accounts_schedule.

API Changes​

  • No schema changes. Both operations already documented the 422 response.

MCP​

  • schedulePost and approvePost (and the connector's schedule-post) return the validation error for malformed schedule input instead of an internal error.

MCP connector — 2026-10-14​

No API or spec change; the SDKs are unchanged.

MCP​

  • New connector tool retry-failed-post: publishes a post again to one account whose publishing failed (the API's retryPostAccount). Takes the account id get-post shows, or its uuid.
  • New connector tool get-ai-credits: the GET /{workspaceUuid}/usage/ai payload, so an assistant can check credits before repurpose-content.
  • create-post takes an optional idempotency_key: a retry with the same key and arguments within 24 hours returns the first result and creates no second post, like the API's Idempotency-Key header.

v1.17.1 — 2026-10-14​

What's New​

  • Queue caps on the right day: addPostToQueue and the queue flag on createPost, updatePost and validatePost now check the daily cap on the UTC day of the queue slot the post will get (the earliest next free slot of its accounts' posting schedules), not today. A post queued after today's last slot is no longer refused because today is full. See Guardrails.

API Changes​

  • No schema changes. The addPostToQueue description now says which day the cap is checked on.

MCP​

  • add-post-to-queue checks the cap on the queue slot's day, as above.
  • update-post checks the versions the post holds after the update: the saved per-account versions plus the ones you send. A content-only update no longer reports a duplicate for an account that keeps its own version.
  • schedule-post and add-post-to-queue answer a non-string uuid with the validation error instead of an internal error.
  • Since v1.16.0, update-post on a scheduled post checks it at its scheduled time. Changing its content or accounts is refused when an account needs reconnecting (account_unauthorized) or the day's cap is full. Drafts only get the duplicate check.

v1.16.0 — 2026-10-13​

What's New​

  • Guardrails on schedule and queue: schedulePost (POST /{workspaceUuid}/posts/schedule/{postUuid}) and addPostToQueue (POST /{workspaceUuid}/posts/add-to-queue/{postUuid}) now run the same post guardrails as createPost/updatePost on the saved post. Previously, a draft followed by a schedule call skipped them. Schedule checks at the requested or saved publish time. Queue checks as publishing today (UTC). See Guardrails.

API Changes​

  • schedulePost and addPostToQueue add 409 (PostConflict, duplicate content; override with X-Allow-Duplicate: true) and 429 (DailyCapReached, daily cap or X monthly allowance, e.g. x_monthly_cap_reached). Their 422 is now PostUnprocessable (the account cannot publish, plus the existing validation errors).
  • Scheduling a post over the X monthly cap now returns 429 x_monthly_cap_reached from the guardrail, where it used to return the engine's 422 quota response.

MCP​

  • The MCP connector's create-post, update-post, schedule-post and add-post-to-queue tools enforce the same guardrails (duplicate guard with allow_duplicate, daily caps, account health, X monthly allowance) and return a refusal as a tool error carrying the code.

v1.15.0 — 2026-10-12​

What's New​

  • Webhook envelope: every delivery now carries a delivery id (repeated on retries, so you can de-duplicate), occurred_at and workspace_uuid next to event and data. See Webhooks.
  • Platform post IDs: Account.provider_post_id (the network's own ID of the published post) is returned wherever an account is embedded in a post, including webhook payloads.
  • Content Credentials on every upload: a JPEG that carries C2PA Content Credentials and needs no resize is now stored byte for byte on uploadMedia and the dashboard too, not only on uploadMediaBase64. See AI disclosure.
  • PDFs on more upload paths: the chunked upload (initiateChunkedUpload) accepts PDFs up to 100 MB, and the remote-URL import (initiateRemoteUpload) accepts a PDF under 10 MB. See Media uploads.

API Changes​

  • Account gains provider_post_id (nullable string).

MCP​

  • The MCP connector grew from 24 to 40 tools: validate-post (dry run with guardrails), get-account-health (with X usage), brand voice get/update, upload-media (base64), list-automations, create-recycling-rule, add-rss-feed, create-report, repurpose-content, agents and their approval queue (list-agents, list-agent-proposals, approve-agent-proposal, reject-agent-proposal) and inbox drafts (list-inbox-drafts, create-inbox-draft; approving stays in the app). These run the REST API in-process, so scoped keys, guardrails and the activity log (client: mcp-connector) apply.
  • Scoped API keys are refused by the connector (403 insufficient_scope); use OAuth, an unscoped dashboard token, or the advanced server.

v1.14.0 — 2026-10-12​

What's New​

  • AI labels on Instagram and YouTube: ai_generated now sets Instagram's "AI info" label and YouTube's synthetic-media disclosure as well as TikTok's. X-AI-Disclosure is applied for all three. See AI disclosure.
  • LinkedIn document posts: upload a PDF (uploadMedia, uploadMediaBase64) and attach it as the only media item of a LinkedIn post to publish it as a document carousel. See Media uploads.

API Changes​

  • PostVersion.options documents youtube (PostOptionsYoutube, with contains_synthetic_media) and instagram / instagram_standalone (PostOptionsInstagram, with ai_generated).
  • createPost / updatePost return 422 when a PDF is attached next to other media or to non-LinkedIn accounts.

v1.13.0 — 2026-10-12​

What's New​

  • AI usage endpoint: GET /{workspaceUuid}/usage/ai returns text and image credits used and remaining in the current cycle. Image generation now draws from its own credit pool (429 image_credits_exhausted when it runs out).
  • Learn brand voice from past posts: POST /{workspaceUuid}/brand-voice/learn builds the brand-voice profile from the workspace's published posts.
  • Pagination (page, per_page) on the template, hashtag-group and account-group lists; list totals respect account-limited API keys.

v1.12.0 — 2026-10-11​

What's New​

  • Webhook subscriptions API: n8n, Zapier and Make triggers can register and remove their own webhooks. See Webhooks.

API Changes​

New endpoints

MethodEndpointDescription
GET POST/{workspaceUuid}/webhooks/subscriptionsList webhooks, subscribe a callback URL
DELETE/{workspaceUuid}/webhooks/subscriptions/{subscriptionUuid}Unsubscribe

v1.11.0 — 2026-10-11​

Includes spec 1.10.0, published the same day.

What's New​

  • Templates, hashtag groups, account groups, posting schedule and competitors are now in the API (24 operations). See Core Concepts.
  • Inbox API (rolling out): agents read comments and draft replies; a person approves every reply before it is posted. Bluesky, Mastodon, LinkedIn Pages and Google Business Profile first. See Inbox.

API Changes​

New endpoints

MethodEndpointDescription
GET POST/{workspaceUuid}/templatesList and create templates
GET PUT DELETE/{workspaceUuid}/templates/{templateId}Read, update, delete a template
GET POST/{workspaceUuid}/hashtag-groupsList and create hashtag groups
GET PUT DELETE/{workspaceUuid}/hashtag-groups/{hashtagGroupId}Read, update, delete a hashtag group
GET POST/{workspaceUuid}/account-groupsList and create account groups
GET PUT DELETE/{workspaceUuid}/account-groups/{accountGroupUuid}Read, update, delete an account group
GET PUT/{workspaceUuid}/posting-scheduleRead and set queue times
DELETE/{workspaceUuid}/posting-schedule/{accountId}Reset an account to the workspace schedule
GET/{workspaceUuid}/posting-schedule/recommendedRecommended posting times
GET/{workspaceUuid}/queue/slotsUpcoming queue slots
GET POST/{workspaceUuid}/competitorsList and track Instagram competitors
GET DELETE/{workspaceUuid}/competitors/{competitorId}Competitor detail, stop tracking
GET GET PATCH/{workspaceUuid}/inbox/conversations[/{conversationId}]Inbox conversations
POST/{workspaceUuid}/inbox/conversations/{conversationId}/draftsDraft a reply
GET PUT DELETE/{workspaceUuid}/inbox/drafts[/{draftId}]Reply drafts
POST/{workspaceUuid}/inbox/drafts/{draftId}/approve, /rejectReview a draft (publish ability)

v1.9.0 — 2026-10-11​

What's New​

  • Idempotent requests. Send Idempotency-Key with any POST; a retry with the same key returns the original response (Idempotent-Replayed: true) and creates nothing.
  • Agent guardrails. Create and update requests from the API are refused when they duplicate a post from the last 24 hours (409 duplicate_post), exceed an account's daily post cap (429 account_daily_cap_reached), or schedule to an account that must be reconnected (422 account_unauthorized). See Agent Guardrails.
  • Readable platform errors. A failed account's errors now starts with a readable line such as "Facebook: (#200) Permissions error (code 200)", followed by {platform, code, message, raw} with the network's original response.
  • Webhooks: deliveries carry X-PostBoost-Workspace, and the webhook docs now describe the real events, payload, signature and retry schedule.

API Changes​

New endpoints

MethodEndpointDescription
POST/app/api/{workspaceUuid}/posts/validateDry run of create-post: validation, network requirements, guardrails
GET/app/api/{workspaceUuid}/account-healthAuthorized / token / daily-cap usage per account
GET/app/api/{workspaceUuid}/activityLog of API write requests and their outcomes

Changed

  • POST /posts and PUT /posts/{postUuid} can return 409, 429 and new 422 codes (see above); allow_duplicate added to the post body.
  • Breaking for raw error parsing: a raw network error in an account's errors moved from errors.error.message to errors[1].raw.error.message; errors[0] is the readable line.
  • Account objects inside posts document external_url and errors.

MCP​

  • New curated MCP connector at https://postboost.co/mcp: 24 workflow tools with titles and read-only / destructive / idempotent / open-world annotations, signed in with OAuth 2.1 (PKCE, dynamic client registration) from Claude.ai, ChatGPT, Claude Code, Cursor and VS Code. Dashboard access tokens still work for headless use.
  • The spec-generated server at /mcp/postboost is now the advanced option (one tool per API operation) and gained validatePost, getAccountHealth and listApiActivity from API 1.9.0.

Docs​


v1.8.0 — 2026-10-11​

What's New​

  • Major platform upgrade: new scheduling, analytics and media engine underneath the API.
  • Analytics over the API: a cross-platform workspace summary, per-account analytics, a posting-times heatmap, best-time recommendations, and per-post analytics.
  • Media folders: organize the media library in nested folders, and file uploads straight into a folder.
  • Mention book: store people and pages once and reference them in post bodies with {{mention:<reference>}}. Each mention resolves to the right handle on each network when the post is published.
  • Retry publishing to a single failed account without re-sending the post to accounts that already published.
  • Per-account scheduling: give each account its own departure time with accounts_schedule.
  • Post dates and times are now read in the request's timezone. date, time and accounts_schedule use the timezone you send, and fall back to the workspace timezone. PUT /posts/{postUuid} used to ignore timezone and now honours it.
  • A second MCP endpoint is available at /app/mcp (30 tools), next to the existing /mcp/postboost.
  • The Engagement inbox that ships with this release is not enabled yet.

API Changes​

New endpoints

MethodEndpointDescription
GET/app/api/workspacesList the workspaces the authenticated user belongs to, with their role in each
GET/app/api/{workspaceUuid}/analyticsCross-platform analytics summary (period or date_from/date_to)
GET/app/api/{workspaceUuid}/analytics/accounts/{accountUuid}Analytics for one account (type, page, per_page, sort_by, sort_dir)
GET/app/api/{workspaceUuid}/analytics/posting-timesEngagement heatmap by weekday and hour (accounts[], timezone)
GET/app/api/{workspaceUuid}/analytics/best-timesNext recommended posting time per account (accounts[], from, timezone)
GET/app/api/{workspaceUuid}/posts/{postUuid}/analyticsEngagement metrics for a published post
POST/app/api/{workspaceUuid}/posts/retry/{postUuid}/{accountUuid}Retry publishing to one failed account
GET/app/api/{workspaceUuid}/media/foldersList the media folder tree
POST/app/api/{workspaceUuid}/media/foldersCreate a media folder
PUT/app/api/{workspaceUuid}/media/folders/{folderUuid}Rename or move a media folder
DELETE/app/api/{workspaceUuid}/media/folders/{folderUuid}Delete a media folder (optionally with its media)
GET/app/api/{workspaceUuid}/mentionsList the mention book (paginated, q search)
POST/app/api/{workspaceUuid}/mentionsCreate a mention
GET/app/api/{workspaceUuid}/mentions/{mentionId}Get a mention
PUT/app/api/{workspaceUuid}/mentions/{mentionId}Update a mention
DELETE/app/api/{workspaceUuid}/mentions/{mentionId}Delete a mention
GET/app/api/panel/workspaces/{workspaceUuid}/ai-creditsAI credit availability and balance for a workspace (admin)

No endpoints were removed.

Deprecations

  • Workspace.hex_color is deprecated: workspaces no longer store a colour (workspaces now use avatar), so panel workspace responses now return the constant 4f46e5 to keep existing clients working. Use avatar (nullable URL) instead. No breaking changes in this release.
  • hex_color in the workspace create/update body is no longer required. It is deprecated and ignored.
  • Every /app/api/{workspaceUuid}/... endpoint now returns 403 when the workspace's plan does not include API access (plan capability api).
  • Scheduling, approving, queueing, and creating or updating a scheduled post now check provider requirements up front (for example a network that requires media). Failures return 422 instead of failing at publish time.
  • POST /app/api/{workspaceUuid}/posts/schedule/{postUuid} with postNow: true now starts publishing within the request.
  • DELETE /app/api/panel/workspaces/{workspaceUuid}/subscription/generic now returns 422 when the workspace has no generic subscription. It used to return success.

New optional request fields

  • Posts (POST/PUT /posts): accounts_schedule[] (account_id, date, time) and versions[].synced_with.
  • POST /posts/schedule/{postUuid} and POST /posts/approve/{postUuid}: date, time, timezone, accounts_schedule[].
  • POST /posts/add-to-queue/{postUuid}: timezone. Each account now gets its own queue slot.
  • GET /accounts: keyword, group, providers[], unauthorized filters.
  • GET /media: limit, folder (UUID or root), mime_types[], type, keyword, usage, favorite, sort.
  • POST /media, POST /media/remote/initiate, POST /media/chunked/{uploadUuid}/complete: folder.
  • PUT /media/{mediaUuid}: name and folder. alt_text is no longer required.
  • DELETE /media: items[] accepts media UUIDs as well as numeric IDs (up to 500).
  • POST /panel/workspaces/{workspaceUuid}/subscription/cancel: immediately. Past-due subscriptions can now be cancelled too.

New response fields (additive)

  • Account: profile_url. When embedded in a post, it also has publish_status and published_parts.
  • Media: is_processing, width, height, duration_ms, size, alt_text, folder.
  • Post: publish_progress (total, published, failed), present only while publishing. status is now documented with all of its values: draft, scheduled, published, failed, needs_approval, publishing.
  • Post version: synced_with.
  • Panel user: avatar.
  • Schedule/queue responses: needs_approval.

Platform & Media​

  • Video uploads are converted in the background, so a fresh upload may return is_processing: true until conversion finishes.
  • Media can be filed in nested folders. Deleting a folder moves its contents up a level unless delete_media is set.

v1.7.0 — 2026-04-12​

API Changes​

  • AI image endpoints (/ai/image-generate, /ai/image-edit, /ai/image-variations) return the standard Media schema instead of GeneratedImageItem, so generated images can be attached to posts directly. SDKs regenerated.

v1.6.0 — 2026-04-11​

New API Endpoints​

AI — Image Generation & Editing

MethodEndpointDescription
POST/{workspaceUuid}/ai/image-promptBuild an optimized image generation prompt from a social media caption and platform. Use this to preview the prompt before calling image-generate. Costs 1 AI credit per call.
POST/{workspaceUuid}/ai/image-generateGenerate one or more social media images from a caption (standard mode) or a raw prompt (developer mode). Images are saved directly to the workspace media library. Credits: count × credit_weight (default 5 per image).
POST/{workspaceUuid}/ai/image-editEdit an existing workspace media image using a text prompt and optional mask. Results are saved to the media library. Credits: count × credit_weight.
POST/{workspaceUuid}/ai/image-variationsGenerate one or more variations of an existing workspace media image. Results are saved to the media library. Credits: count × credit_weight.
POST/{workspaceUuid}/ai/image-alt-textAnalyze an existing workspace media image and generate accessible alt text. The alt text is saved back to the media record. Costs 1 AI credit per call.

No breaking API changes.​


v1.5.0 — 2026-04-04​

New API Endpoint​

AI

MethodEndpointDescription
POST/{workspaceUuid}/ai/blog-to-socialGenerate platform-specific social media captions from a blog post URL or title and excerpt. Supports tone, hashtags, CTA, language, and optional draft post creation with image import. Consumes one AI credit per platform.

No breaking API changes.​


v1.4.0 — 2026-03-23​

First SDK releases: Python, Node, PHP, Go, Ruby, C#, Rust, Java and the CLI were published as 1.0.0 to 1.4.0 on this day.

Platform Improvements​

  • Added Mastodon service management endpoints
  • Facebook API updated to v25.0
  • Improved HEIC and HEIF image upload support
  • Post status is now set to DRAFT when moved to trash

Media​

  • Chunked upload support for large files — POST /{workspaceUuid}/media/chunked/initiate
  • Remote media download from URL — POST /{workspaceUuid}/media/remote/initiate
  • Compatible with more video formats (MP4, MOV, AVI, WebM)
  • Additional image format support
  • Resolved infinite loading issue in Media Library

No breaking API changes.​


v1.1.0 — 2025-12-29​

New API Endpoints​

Workspaces (admin token required)

MethodEndpointDescription
GET/panel/workspacesList all workspaces
POST/panel/workspacesCreate workspace
GET/panel/workspaces/{uuid}Get workspace
PUT/panel/workspaces/{uuid}Update workspace
DELETE/panel/workspaces/{uuid}Delete workspace
DELETE/panel/workspacesBulk delete workspaces
POST/panel/workspaces/{uuid}/usersAdd user to workspace
PUT/panel/workspaces/{uuid}/usersUpdate user role
DELETE/panel/workspaces/{uuid}/usersRemove user from workspace

Subscriptions (admin token required)

MethodEndpointDescription
GET/panel/workspaces/{uuid}/subscriptionGet subscription
POST/panel/workspaces/{uuid}/subscriptionCreate subscription
PUT/panel/workspaces/{uuid}/subscriptionUpdate subscription
DELETE/panel/workspaces/{uuid}/subscriptionDelete subscription
POST/panel/workspaces/{uuid}/subscription/newCheckout new subscription
PUT/panel/workspaces/{uuid}/subscription/change-planChange plan
POST/panel/workspaces/{uuid}/subscription/cancelCancel subscription
POST/panel/workspaces/{uuid}/subscription/resumeResume subscription
POST/panel/workspaces/{uuid}/subscription/genericAdd generic subscription
DELETE/panel/workspaces/{uuid}/subscription/genericRemove generic subscription

Users (admin token required)

MethodEndpointDescription
GET/panel/usersList users
POST/panel/usersCreate user
GET/panel/users/{id}Get user
PUT/panel/users/{id}Update user
DELETE/panel/users/{id}Delete user
DELETE/panel/usersBulk delete users

Receipts (admin token required)

MethodEndpointDescription
GET/panel/receiptsList receipts
POST/panel/receiptsCreate receipt
GET/panel/receipts/{id}Get receipt
PUT/panel/receipts/{id}Update receipt
DELETE/panel/receipts/{id}Delete receipt
DELETE/panel/receiptsBulk delete receipts

v1.0.0 — 2025-08-15​

Initial Release​

Workspace endpoints (workspace API token required)

MethodEndpointDescription
GET/{workspaceUuid}/accountsList connected social accounts
GET/{workspaceUuid}/accounts/{id}Get account
GET/{workspaceUuid}/postsList posts
POST/{workspaceUuid}/postsCreate post
GET/{workspaceUuid}/posts/{uuid}Get post
PUT/{workspaceUuid}/posts/{uuid}Update post
DELETE/{workspaceUuid}/posts/{uuid}Delete post
DELETE/{workspaceUuid}/postsBulk delete posts
POST/{workspaceUuid}/posts/schedule/{uuid}Schedule post
POST/{workspaceUuid}/posts/add-to-queue/{uuid}Add to queue
POST/{workspaceUuid}/posts/approve/{uuid}Approve post
GET/{workspaceUuid}/mediaList media
POST/{workspaceUuid}/mediaUpload media
GET/{workspaceUuid}/media/{id}Get media
PUT/{workspaceUuid}/media/{id}Update media
DELETE/{workspaceUuid}/mediaDelete media
GET/{workspaceUuid}/tagsList tags
POST/{workspaceUuid}/tagsCreate tag
GET/{workspaceUuid}/tags/{id}Get tag
PUT/{workspaceUuid}/tags/{id}Update tag
DELETE/{workspaceUuid}/tags/{id}Delete tag