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

# How do I page older or newer Telegram messages, and can search use before?

> Paginate telegram.web.fetch_channel_posts with unchanged before_cursor or after_cursor integers, and filter telegram.web.fetch_channel_search locally because channel search has no cursor or date parameters.

To page a Telegram channel with `telegram.web.fetch_channel_posts`, pass the integer cursor from the previous response unchanged: `before` reads older messages and `after` reads newer messages. Channel search with `telegram.web.fetch_channel_search` takes no `before`, `after` or date input; to cover a time window, read a bounded set of messages and filter the messages by their returned timestamps.

## How do `before` and `after` differ?

In `telegram.web.fetch_channel_posts`, `before` requests older messages and `after` requests newer messages. Omit both `before` and `after` on the first request. On the next request, pass `data.pagination.before_cursor` or `data.pagination.after_cursor` from the previous response, unchanged, for the direction you want. Both cursors are integers. Do not increment a cursor as a page number, and do not send `before` and `after` in the same request.

## Does `limit=100` mean one page of 100?

No. `limit` accepts 1–100 and defaults to 20. Setting `limit` above 20 causes `telegram.web.fetch_channel_posts` to page automatically toward older messages, so `limit=100` does not guarantee a single page of 100 posts. Start with `limit` 20 for a bounded request and save the request and the raw response. Reaching the `limit` does not mean the channel history is exhausted.

## How does channel search work?

`telegram.web.fetch_channel_search` requires `channel` and `query`, and accepts `limit` from 1 to 100. Channel search has no `before`, `after` or date input, so do not copy list-pagination fields into a search request. Channel search covers the named channel only, not all of Telegram. For a time window, collect a bounded set of messages, filter the returned timestamps locally, and report the scanned range rather than claiming complete coverage.

## When should paging stop?

Stop paging when the cursor is missing or repeated, when a response fails, or when you reach your request limit, and save the scanned range and the stop reason. When reading older messages, record the last usable `before_cursor` as your checkpoint. Archive every original response, read the full message objects from `data.messages`, and check each message object for the text and message ID.

For example, you saved the latest batch on Monday. To look for last week's announcements, request older messages with Monday's `before_cursor`. To look for announcements published by Tuesday, request newer messages with Monday's `after_cursor`. Save the older and newer sets separately, then check for overlap by message ID within the channel.

[Complete archive workflow](/guides/telegram-channel-archive)