Skip to content

Scribe Integrations

Prerequisites

This guide assumes you've completed the Quick Start, have a working summon config check, and have enabled the Scribe.

The Scribe agent can monitor external data sources and surface important information to your Slack channel. Each integration is optional — enable whichever ones are useful for your workflow.

Integration What it provides Extra required
Google Workspace Gmail, Calendar, Drive monitoring summon-claude[google]
Slack Browser Monitoring External Slack workspace monitoring summon-claude[slack-browser]

Google Workspace

The Scribe can monitor Gmail, Google Calendar, and Google Drive for important updates.

Setup

Install the Google extra if you haven't already:

uv pip install 'summon-claude[google]'

pipx inject summon-claude workspace-mcp

Run the guided setup to create Google OAuth credentials:

summon auth google setup

The setup is an interactive wizard with a progress roadmap — each step gets a clean screen showing where you are and what's next:

  1. Google Cloud Project — select an existing project or create a new one. If gcloud is installed, the wizard detects your current project and offers it as the default, and can create new projects for you. Console links route through Google's account chooser for multi-account users.
  2. Enable APIs — enable Gmail, Calendar, and Drive APIs. If gcloud is installed, offers to run the command directly; otherwise provides browser links with an option to open them automatically.
  3. OAuth Consent Screen — configure branding and set publishing status to Production (avoids 7-day token expiry). Links open directly in your browser.
  4. Create OAuth Client — create a "Desktop app" OAuth client and download client_secret.json. The wizard accepts either the JSON file path or a manually pasted Client ID + Secret.

Already have a GCP project with OAuth configured?

In step 1, choose "Skip this step" to proceed directly to API enablement and credentials.

Then authenticate with Google:

summon auth google login

This prompts which services need write access (all are read-only by default), then opens a browser for OAuth consent. Credentials are stored in summon's config directory (google-credentials/).

To re-run later and change scope access (e.g., grant or revoke write access for a service), run summon auth google login again — prompt defaults match your current grants so you won't accidentally downgrade.

To verify authentication status:

summon auth google status

This shows whether credentials exist, which scopes are granted (read-only vs read-write per service), and whether the token is still valid.

Enabling the Google collector

The Google collector is auto-detected: when workspace-mcp is installed and Google credentials exist, the scribe automatically uses Google tools. No manual config flag is needed.

To explicitly disable it: summon config set SUMMON_SCRIBE_GOOGLE_ENABLED false

Available Google services are auto-detected from the OAuth scopes granted during summon auth google login. The scribe automatically monitors whichever services the credential supports (e.g., Gmail, Calendar, Drive).

Slack Browser Monitoring

The Scribe can monitor an external Slack workspace using browser-based WebSocket interception. This is separate from the native Slack bot integration used for session interaction — it watches a different workspace (e.g., your company's Slack) via a real browser session.

Browser-based monitoring

Slack channel monitoring uses Playwright to capture WebSocket frames from your Slack workspace. This requires the slack-browser extra and a Chromium-based browser installed on the host.

Setup

Install the Slack browser extra if you haven't already:

uv pip install 'summon-claude[slack-browser]'

pipx inject summon-claude playwright

Authenticate with a Slack workspace

summon auth slack login myteam

This opens a visible browser window at your Slack workspace. Log in normally — the browser closes automatically after detecting your session. Auth state (cookies and localStorage) is saved to summon's data directory.

The WORKSPACE argument accepts:

  • A workspace name: myteam (becomes https://myteam.slack.com)
  • An enterprise name: acme.enterprise (becomes https://acme.enterprise.slack.com)
  • A full URL: https://myteam.slack.com

After login, the command prompts you to select which channels to monitor using an interactive picker.

Enterprise Grid workspaces

Enterprise Grid workspaces serve a workspace picker at their enterprise URL. The scribe handles this automatically by extracting team IDs from the saved browser state and navigating directly to app.slack.com/client/{TEAM_ID}.

Select monitored channels

To change which channels are monitored without re-authenticating:

summon auth slack channels

This uses the cached channel list from the last authentication. To refresh the channel list from Slack:

summon auth slack channels --refresh

Check auth status

summon auth slack status

Shows the configured workspace URL, user ID, auth state age, and monitored channels.

Remove auth state

summon auth slack logout

Removes saved browser auth state and workspace config. This cannot be undone.

Enabling the Slack collector

Once authenticated, the Slack collector auto-enables on the next summon project up. Optionally configure monitored channels and browser:

summon config set SUMMON_SCRIBE_SLACK_BROWSER chrome
summon config set SUMMON_SCRIBE_SLACK_MONITORED_CHANNELS C01ABC123,C02DEF456

DMs and @mentions are always captured regardless of the channel list. The channel list controls which channels have all messages monitored.

How it works

The browser user must be a member of any channel being monitored — the WebSocket only delivers messages for channels the authenticated user belongs to.

The primary auth cookie (d) has a roughly 1-year TTL, so re-authentication is rarely needed. The x cookie (CSRF) is not required.

See also