Integration · Open by default
Any MCP server, no custom code.
If a service speaks MCP, Brigham.ai can use it. The connection is governed like every other one.
Adding a connection
1
Add the server
Point Brigham.ai at the MCP server and sign in.
2
Scope it
Choose which tools are allowed and for which agents.
3
Classify actions
Each tool is read, reversible write or sensitive.
4
Audit it
Every call is checked by policy and logged.
Scopes and access
- Connects as
- A person, or the org if shared
- Access
- Read-only by default
- Reads
- Whatever the server exposes and you allow
- Writes
- Per tool, as a separate grant
- Needs approval
- Tools classified as sensitive
- Visibility
- User-private or org-shared, set at connection
- Availability
- Available
Yours to connect, yours to revoke.
Each person connects their own accounts from the Connections page. Every connection can be revoked on its own, and every connection starts read-only. A write scope is a separate, explicit grant.
- Credentials are encrypted at rest and referenced by id. They never appear in model context, logs or tool results.
- Each agent gets its own connection grants.
- Tool results from outside sources are untrusted data. They cannot approve anything.
Ask it
- ›“Which of my connections are read-only?”
- ›“Use the support tool to find open tickets for Northwind.”
- ›“What did that connection do yesterday?”
Questions, answered.
Is an MCP tool result trusted?
No. Results are untrusted data and cannot approve an action.
Does it skip approvals?
No. Sensitive tools wait for approval like any other action.
Keep reading.
CredentialsBrigham.ai encrypts credentials at rest and refers to them by id. Keys never appear in the model’s context, the logs or a tool result.Read →
Action classesBrigham.ai sorts every action into read, reversible write or sensitive. Sensitive actions always wait for a person.Read →
Untrusted contentEmail, web pages, documents and notes are labelled as data. Nothing that arrives through a tool can approve an action, change a policy or widen a scope.Read →
GitHubBrigham.ai reads your repositories, issues, branches and pull requests, so engineering questions get answers from the source. Read-only by default.Read →