Data Sources
50+ pre-built connectors for the systems you already run. Plus a custom-connector framework for everything else. All sync continuously, respect permissions, and feed directly into the [knowledge base](/docs/platform/knowledge-base).
Connector catalog
🛠 In active development — Q3–Q4 2026. The connector framework, sync model, and permission inheritance are available today. Specific connectors marked with a quarter (Q3 2026 / Q4 2026) below are ready in that release.
Collaboration & chat
| Source | What we ingest | Auth |
|---|---|---|
| Slack | Channels, DMs (with consent), threads, file attachments | OAuth 2.0 |
| Microsoft Teams | Channels, chats, meeting transcripts, attachments | OAuth 2.0 |
| Google Chat | Spaces, conversations, attachments | OAuth 2.0 |
| Discord | Servers, channels, threads | Bot token |
| Mattermost | Teams, channels, threads | Personal access token |
| Source | What we ingest | Auth |
|---|---|---|
| Gmail | Inbox, labels, threads, attachments | OAuth 2.0 |
| Outlook / Exchange | Inbox, folders, threads, attachments, calendar | OAuth 2.0 / Basic auth |
| IMAP (any) | Folders, messages, attachments | App password |
| Fastmail, Zimbra, ProtonMail | Standard mail content | Various |
Cloud storage
| Source | What we ingest | Auth |
|---|---|---|
| Google Drive | Files, folders, shared drives, comments | OAuth 2.0 |
| OneDrive / SharePoint | Files, libraries, lists, metadata | OAuth 2.0 |
| Dropbox | Files, folders, shared spaces | OAuth 2.0 |
| Box | Files, folders, comments, metadata | OAuth 2.0 / JWT |
| iCloud Drive | Files, folders | App-specific password |
| S3 / MinIO / Wasabi | Buckets and prefixes | Access key |
Wikis & documentation
| Source | What we ingest |
|---|---|
| Notion | Pages, databases, blocks, comments |
| Confluence | Spaces, pages, attachments, comments |
| Obsidian | Vaults via local sync or Git |
| Coda | Docs, tables, formulas |
| Outline | Collections, documents, revisions |
Productivity & project
| Source | What we ingest |
|---|---|
| Google Workspace | Docs, Sheets, Slides, Forms |
| Microsoft 365 | Word, Excel, PowerPoint, Forms |
| Trello / Asana / Jira | Boards, cards, tickets, comments |
| Linear | Issues, projects, cycles |
Databases & analytics
| Source | What we ingest |
|---|---|
| PostgreSQL / MySQL / MariaDB | Tables, views, with row-level filters |
| MongoDB | Collections, documents |
| Elasticsearch / OpenSearch | Indexes, queries |
| Snowflake / BigQuery / Redshift | Tables, views, scheduled queries |
| ClickHouse | Tables, materialized views |
CRM & sales
| Source | What we ingest |
|---|---|
| Salesforce | Accounts, contacts, opportunities, cases, custom objects |
| HubSpot | Companies, contacts, deals, tickets |
| Pipedrive, Zoho, Copper | Standard CRM objects |
Domain-specific
| Source | Industry |
|---|---|
| SAP (S/4HANA, ERP, PM) | Manufacturing, finance, HR |
| iManage / NetDocuments | Legal document management |
| Epic FHIR / Cerner | Healthcare EHR |
| OSIsoft PI (PI Web API) | Industrial historians |
| SCADA via OPC UA | Industrial control |
| Bloomberg Terminal API | Finance market data |
Identity & metadata
- SAML 2.0 (Okta, Azure AD, Auth0, Keycloak, Ping)
- OIDC
- SCIM 2.0 for user/group provisioning
- LDAP / Active Directory
Custom connectors
For sources we don’t pre-support, build a custom connector:
- REST — OpenAPI spec, define the endpoints to poll, the auth, the response schema
- GraphQL — same idea with a query template
- gRPC — proto file + connection details
- Webhook — receive push notifications, transform, ingest
- Python script — bring your own runtime for arbitrary logic
Custom connectors are versioned and run in isolated containers. Schedule, retry policy, and error handling are all configurable.
Sync model
Each connector has a sync cadence:
- Continuous — webhook-driven where supported (Slack, Teams, GitHub)
- Polled — every 1–60 minutes, configurable
- Triggered — only on demand
- Bulk + delta — initial bulk ingest, then delta sync
Permissions inheritance
Most connectors carry through source-system permissions. A document a user can’t read in Google Drive is a document they can’t read in StellarBase, even if they have Base access. This is configured per-connector — either via the source’s native ACL (Drive, SharePoint), or via a mapping you define.
Rate limits & backoff
Every connector respects source rate limits with exponential backoff. We never get your account banned for excessive requests. Per-connector throttling is configurable.
Error handling
Failed documents are logged with the failure reason. The connector continues processing other documents. You can review failures in the Base UI and retry individually or in bulk.
Related
- Knowledge Base — what happens after ingest
- Supported Formats — file types we parse
- Auth & RBAC — permission inheritance details