Tool connectors
Connect Notion, Slack, Linear, and GitHub so Sundial's agents can read context from them.
Sundial's agents answer best when they can see the context that lives in the tools your team already uses. Tool connectors let you grant Sundial read access to those tools over the Model Context Protocol (MCP), so an agent can pull in the relevant details as it works.
Available connectors
| Connector | What agents can read |
|---|---|
| Notion | Pages and content from your Notion workspace |
| Slack | Channels, messages, and files (also powers asking @sundial in Slack) |
| Linear | Issues and project metadata |
| GitHub | Repository metadata and code |
Connecting a tool
Connect a tool from the Context Engine. Each connector uses OAuth: you authorize Sundial against the tool, and access is scoped to what you approve. You can revoke a connection at any time.
Read-only by design
Tool connectors are read-only. Agents can pull context in to inform an analysis, but they don't write back to your Notion, Slack, Linear, or GitHub through these connectors. This keeps the blast radius small: connecting a tool can only ever surface information, never change it.
GitHub for content sync
The GitHub connector here is for reading repository context. Syncing Sundial's own AI-managed context (semantic models, playbooks, evals) through your repository is a separate setup; see the GitHub integration under Connecting your data.
More connectors are on the way. If you need a tool that isn't listed, let your Sundial contact know.