API and MCP Logs
See every API call and MCP tool call made on your PostEverywhere account: who made it, what it did, how long it took and why it failed. Kept for 30 days on every plan.
Every call made to the PostEverywhere API with an API key is logged, including the calls your AI assistant makes through the MCP server. Open Logs in the dashboard (it sits just under Developers in the sidebar) to see what Claude, ChatGPT, Cursor, n8n or your own code did, when, and whether it worked.
What gets logged
Every public API call (/v1) made with an API key. That covers:
- AI assistants over MCP. Connecting Claude, ChatGPT or Cursor creates a key, so every tool call they make is logged, with the tool name (for example
create_post). - Your own code, the SDK and the CLI.
- Automation tools such as n8n, Zapier and Make.
Not logged: webhook deliveries, the account-connect (OAuth) screens, activity in the web app or the iPhone app, and requests rejected because the key was missing or invalid.
What each entry shows
| Field | What it tells you |
|---|---|
| Status | The HTTP status code, so failures stand out |
| Call | The action in plain words ("Create post"), plus the method and path |
| Made by | Claude, ChatGPT, n8n or API, and the name of the key |
| MCP tool | The tool the assistant called, when the call came through MCP |
| Time | How long the call took |
| When | When it happened |
| Request ID | The ID that matches the X-Request-Id response header |
| Error | The error code and message, when the call failed |
| Request and response | The bodies, with secrets redacted (see below) |
Click any row to open the full detail in a side panel.
Finding a call
- Time range: last hour, 24 hours, 7 days or 30 days.
- Outcome: all calls, succeeded only or failed only.
- Who: one key or one assistant.
- Search: a path, an MCP tool name, an error code or a request ID.
The totals at the top show how many calls were made in the range, how many failed (and what share), and the typical response time. Each call is one row, and 100 calls load at a time; use Load older calls to go further back. The list refreshes when you press Refresh; it doesn't stream live.
Matching a log entry to your code
Every /v1 response carries an X-Request-Id header. The same value is in meta.request_id in the response body and on the log entry, so you can paste it into the search box and jump straight to the call. Include it when you contact support about a failed request. Error responses are described in Errors.
Who can see logs
Owners and admins see every call across the organisation. Members see only the calls made by the keys and assistants they connected themselves.
How long logs are kept
30 days, on every plan. There is no export.
What is redacted
Passwords, tokens, API keys, webhook URLs and other secret-looking values are replaced with [redacted] before an entry is saved, and file data isn't stored. Post captions are kept, so you can see exactly what an assistant wrote. For successful read requests, only the status and timing are stored, not the response.
What logs don't tell you
A log entry shows PostEverywhere's answer to the call. It doesn't show what happened later on the social platform: a post that was accepted and scheduled can still fail when it publishes. To see whether a post went live, check post results or subscribe to webhooks. Logs can't be read through the API or the MCP server, so an assistant can't check its own log.
Related: Launch post: AI agent logs · Errors · Rate limits · MCP guide · Webhooks · Testing your integration