Feeds connect your indexes to external document sources and keep them synchronized. Instead of manually uploading documents, you can set up a feed that automatically imports and updates content from your existing systems.Documentation Index
Fetch the complete documentation index at: https://docs.circuit.ai/llms.txt
Use this file to discover all available pages before exploring further.
Supported sources
Circuit integrates with a broad and growing set of platforms, including:- Collaboration platforms: Confluence, Notion, SharePoint
- Cloud storage: Google Drive, Microsoft OneDrive, Dropbox, Box
- Enterprise systems: ERPs, CRMs, PLM systems, and other line-of-business applications
- Content management: web CMS platforms, knowledge bases, and wikis
How integrations work
When you connect a feed to an index:- Initial import: Circuit imports all documents from the configured source
- Ongoing updates: Circuit periodically checks for changes and re-imports updated documents
- Deletions: if a document is removed from the source, it is removed from the index on the next update
Setting up an integration
Choose the source
Navigate to your index settings and select the feed type for your source platform.
Authenticate
Connect Circuit to your external system using OAuth or an API token, depending on the source.
Select content
Choose which spaces, folders, or pages to import. You can sync an entire workspace or select specific areas.
Connecting OneDrive
Microsoft OneDrive connects through delegated OAuth. From an index’s Integrations tab, choose New integration and click Import from OneDrive. Microsoft prompts you to sign in and consent to the permissions below. After consent, Circuit opens the OneDrive picker so you can select the specific files and folders to sync into the index.Microsoft Graph permissions
The Microsoft consent screen asks for the following Microsoft Graph scopes:Files.Read: read the content and metadata of your files so Circuit can import them.Files.Read.All: let you pick shared files and folders you have access to across your organization, not just ones you own.Files.ReadWrite: granted so Circuit’s feed can access the files it needs to import. Circuit only reads your OneDrive content. It does not create, modify, or delete files through this integration.offline_access: issues a refresh token so Circuit can keep the feed in sync in the background without asking you to reauthenticate each time the short-lived access token expires (roughly every hour by default).
Admin consent in restricted tenants
If your organization uses Microsoft Entra ID (formerly Azure AD) with policies that require admin approval for third-party applications, users will be blocked at the consent screen until an administrator pre-approves Circuit for these scopes. Ask your Entra ID administrator to grant tenant-wide admin consent for the Circuit enterprise application before rolling the integration out to users.Circuit only reads from the files and folders you select in the picker. Files outside the selected scope are not imported, even though the granted permissions technically allow broader access.
Best practices
- Be selective: don’t sync your entire cloud storage. Choose the folders that contain relevant, current documentation.
- Organize at the source: feeds work best when your external content is already well-organized. A clean source structure produces a clean index.
- Monitor feed health: check periodically that feeds are syncing successfully. Authentication tokens can expire and need to be refreshed.
- Avoid duplicates: if the same document exists in multiple feeds or is also manually uploaded, you’ll get duplicate content in the index, which can confuse agents.
Feed update frequency depends on your plan and configuration. Changes in the source typically appear in Circuit within a few hours.