One unified social media API for developers and AI agents

A social media API lets software publish, schedule and read social media content without anyone clicking through each app. PostEverywhere's social media API is one REST API for all top platforms, so you integrate once instead of building a separate connection to Instagram, TikTok, LinkedIn, YouTube and the rest. It handles OAuth, token refresh, rate limits, retries and per-platform formatting for you.

All top platforms: Instagram, TikTok, LinkedIn, YouTube, Facebook, X (Twitter), Threads, Pinterest, Bluesky, Discord, Telegram, WordPress and Google Business Profile. Every post goes out through the platform's official API, and the platform list on this page shows which API each one uses. Facebook posts go to Pages, Instagram needs a Professional account, and Telegram works with channels, groups and supergroups where the bot is an admin.

Text, images, video, carousels, X threads and first comments. Each post carries one video at most; carousels take up to 10 images on Instagram, 1 to 35 photos on TikTok and 2 to 5 images as a Pinterest carousel pin, and LinkedIn takes PDF document posts. Add a first comment with settings.firstComment on Instagram, Facebook, X, Threads and LinkedIn, and post to an X Community with platform_content.x.settings.community_id. Not supported: Instagram collab, location or product tags, LinkedIn polls and articles, Facebook Stories, and comment replies or DMs.

API access is included on every plan, starting with Lite at $9/mo, and there is no separate API fee. Lite already includes API keys and the hosted MCP server; higher plans add connected accounts, AI credits and seats. There is no free plan: every plan starts with a 7-day trial, and a card is required.

Yes. Every call made with an API key, including MCP tool calls from Claude, ChatGPT and Cursor, is listed on the Logs page in the dashboard with its status, duration, who made it and any error. Every response carries an X-Request-Id header that matches the log entry. Logs are kept for 30 days on every plan, with secrets redacted.

There’s no posting cap from PostEverywhere on any plan. Each platform enforces its own limits, and very fast bursts of requests can be rate limited for a moment: the API returns 429 with a Retry-After header. Exact figures for developers are in the [rate limits reference](/docs/rate-limits). Publishing to each platform has built-in retry and rate-limit handling, so a single call reaches every network reliably.

Your own code sends an API key as a Bearer header. An owner or admin creates keys on the Developers page, gives each one read, write or AI permissions, limits it to specific accounts and can revoke it at any time. AI agents such as ChatGPT and Claude connect to the hosted MCP server by signing in to PostEverywhere and approving access, so no key is pasted into a chat.

There are 13 event types: 6 for posts, 2 for accounts, 2 for media and 3 for approvals, including post.published, post.failed, post.partially_failed and account.reconnect_needed. Every delivery is signed with HMAC-SHA256 in the X-PostEverywhere-Signature header, and failed deliveries are retried with exponential backoff. Each organisation can have up to 25 webhook subscriptions.

Open the Developers page in the PostEverywhere dashboard (under Manage in the sidebar) and create a key: name it, choose its permissions (Read, Write and/or AI) and copy it. It starts with pe_live_ and is shown only once, so store it securely. Revoke it from the same page at any time.

Each platform publishes independently. If Instagram succeeds but TikTok fails, for example because a token expired, GET /v1/posts/{id}/results shows exactly which destinations succeeded and which failed. POST /v1/posts/{id}/retry retries only the failed destinations, and POST /v1/posts/retry-failed retries failures in bulk by platform, account or time window.

GET /v1/me returns the API key's context in one call: the organisation, the plan, current quota usage, the scopes granted to the key and feature flags. SDKs use it to configure themselves without hardcoded IDs, and white-label tools use it to show plan-aware options.

There is no separate sandbox. To test safely, connect accounts you use only for testing, or schedule posts far in the future, inspect them through the API and delete them before they publish. Any key can be revoked instantly from the Developers page.

Yes. The API handles platform authentication, media processing and publishing while you build the interface, so agencies and SaaS builders can offer scheduling under their own brand. To publish from a CMS or blog, call POST /v1/posts when your CMS publishes (or use an RSS-to-post workflow in n8n or Zapier), and subscribe to post.published webhooks to close the loop.

Yes. One bulk API call creates up to 50 posts. If you would rather use a spreadsheet, Bulk Upload in the app takes a .csv file of up to 200 posts.

Yes. Install @posteverywhere/sdk for a fully typed Node client with promises and inline documentation. You can also call the REST API directly with a single Bearer token from any language, or run the MCP server locally with npx -y @posteverywhere/mcp.

Yes. Install @posteverywhere/cli with npm install -g @posteverywhere/cli to post and schedule across all top platforms from your terminal. Run posteverywhere login once, connect your accounts, and publish. Every command also supports --json output, so the same CLI works as a tool for AI agents like Claude Code and Cursor.

Yes. POST /v1/media/upload-from-url fetches a public file for you. Images (JPEG, PNG, GIF, WebP, HEIC and HEIF) up to 25 MB are ready straight away; MP4 video up to 4 GB imports in the background, so you check the media item until it is ready. Imported media counts toward your storage.

Yes. POST /v1/posts/{id}/approve approves a pending post, and POST /v1/posts/{id}/request-changes sends it back with a note the author sees. Posts created with an API key skip approval by default; to send one for review, create it with draft: true and request_approval: true. GET /v1/queue and GET /v1/posts?approval=pending list what is waiting. The MCP server can read the queue but cannot approve posts or request changes.

Use the MCP server when an AI agent such as ChatGPT, Claude or Cursor does the work in a conversation: it connects once, sees the tools and calls them in plain language, with nothing to install. Use the REST API or SDK when your own code, a backend job or a no-code tool such as n8n publishes. Both work on the same account and the same posts; see what an agent can do through our MCP server.

Yes. GET /v1/best-times returns the best weekday and hour slots for one account, several accounts (one combined pick) or a whole platform, in the time zone you ask for. Each slot says what it is based on: the account's own posts, what works across PostEverywhere on that platform, or general guidance. next_best is the soonest good slot at least 30 minutes ahead, so pass its next.iso as scheduled_for when you create the post. Through MCP, create_post and schedule_post take schedule_at: "best_time" to do this in one step.

Deleting a published post removes it from PostEverywhere only; it stays live on the platform. The exception is X: add ?delete_on_platforms=x to the delete request and the live X post is deleted too.

Not today. Every automation tool connects through its own generic HTTP block (the n8n HTTP Request node, the Make HTTP module, Webhooks by Zapier, an Airtable script, a Retool REST resource, a Postman collection or the Bubble API Connector) against the REST API. It is one POST request with a Bearer header, so setup takes minutes and nothing needs installing.

Each platform has its own API, OAuth flow, token refresh cycle, rate limits, and media rules. Building and maintaining all of them is weeks of work that keeps breaking as platforms change. PostEverywhere gives you one integration that handles OAuth, token refresh, rate limiting, retries, and per-platform formatting behind a single consistent interface.

Official platform APIs only, including the Meta Graph API, TikTok Content Posting API, and YouTube Data API. There is no fragile browser automation or unofficial scraping, which means reliable delivery and accounts that stay in good standing.