Skip to main content

Slack Integration

Connect Lore to your Slack workspace to capture operational knowledge from thread conversations.

Install the bot​

  1. Open Settings > Integrations in your Lore dashboard
  2. Click Connect Slack
  3. Authorize the Lore bot in your Slack workspace

The bot requires the following OAuth scopes:

  • app_mentions:read — Receive @lore mentions
  • channels:history — Read thread messages for capture
  • chat:write — Post capture confirmations
  • reactions:write — Add status reactions to messages

Commands​

Lore responds to mentions in threads. There are no slash commands — you interact with the bot by mentioning @lore followed by a command:

CommandDescription
@lore startBegin an incremental capture session for the current thread. New messages posted after this point will be captured when you stop.
@lore stopStop the active capture session and ingest all messages posted since @lore start.
@lore captureImmediately capture the entire thread in one shot. No need to start/stop.

Incremental capture (start / stop)​

Use this when a conversation is still happening and you want to capture it as it unfolds:

  1. In the thread, mention @lore start — Lore reacts with 👀 to confirm
  2. Continue the conversation naturally
  3. When the discussion is done, mention @lore stop
  4. Lore ingests the new messages and reports how many documents were saved

If you start a capture and forget to stop it, Lore will automatically close stale sessions when a new @lore start is issued in the same channel.

One-shot capture (capture)​

Use this when a thread is already complete and you want to save the whole thing:

  1. In the thread, mention @lore capture
  2. Lore ingests all non-bot messages in the thread
  3. Re-capturing the same thread updates the existing documents rather than creating duplicates

How it works​

When Lore receives a mention, it:

  1. Verifies the request signature (HMAC-SHA256)
  2. Resolves the workspace from the Slack team ID
  3. Fetches thread messages via the Slack API
  4. Runs the messages through the ingestion pipeline to create or update documents
  5. Posts a confirmation with the document count

Next steps​