> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://apidocs.chattermill.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://apidocs.chattermill.com/_mcp/server.

# Fetch new responses

To pull new feedback out of Chattermill on a schedule:

1. Call [List projects](/api-reference/projects/get-all-projects) to find the key of the project you need.
2. Call [List responses](/api-reference/responses/get-project-responses) for the time window since your last run, and page through every result.

Filter on `updated_from` and `updated_to` rather than `from` and `to`. A response's creation time can be earlier than the moment it reaches Chattermill, for example when feedback is imported in bulk, so a window on creation time can miss it. A response's update time moves forward when it arrives and when Chattermill finishes analysing it.

```bash
curl "https://api.chattermill.com/v1/YOUR_PROJECT/responses?updated_from=2026-11-15T23:55:00Z&updated_to=2026-11-16T01:00:00Z&per_page=100&page=1" \
  -H "Authorization: Bearer YOUR_API_KEY"
```

Keep your jobs reliable with three habits:

* Start each window a few minutes before the previous one ended, so nothing falls between runs.
* Deduplicate on the response `id`, because overlapping windows and later updates return the same response more than once.
* Keep windows short, such as an hour or a day, and request pages until one comes back with fewer results than `per_page`.

If you only want responses Chattermill has finished analysing, add `text_analytics_processed=true`. Because analysis updates the response, it shows up in a later window once it's processed.