API, MCP and org structure
- Who can use it
- Org admin sets up credentials and the hierarchy. Managers read the reports.
- Where
- Admin → MCP, Admin → Hierarchy, Org Hierarchy and Attributes reports
Your delivery data is reachable from your own tools, and it can be shaped like your company. The MCP server lets an AI assistant read your metrics. The REST API serves them to your own scripts. The org hierarchy and attribute slices let you roll numbers up by department, level or location without exposing individuals.
MCP server
The MCP server lets Claude Code or Codex answer questions such as “which team’s review wait grew this month” with your real numbers. It is read-only. It can read, never change.
An org admin sets it up:
- Open Admin → MCP.
- Under MCP credentials, type a Credential name, for example “My Claude Code”.
- Choose Access:
- Acts as me: answers with your role and visibility, like your own sign-in.
- Workspace (aggregates only): team aggregates only. It names nobody.
- Select Create credential and copy it. It is shown once.
- Under Configure your assistants, choose the Claude Code or Codex tab and copy the setup with your credential filled in.
Create one credential per person or tool, so you can revoke each on its own.
The assistant can call these tools: Team metrics, PR workflow, Metric anomalies, Contributor metrics and Pull requests. Contributor-level tools only return people your visibility setting allows, and teams below the minimum group size are withheld.
Full tool list and setup details are in the MCP server reference.
REST API
The REST API serves the same analytics queries to your own scripts and BI tools, with an OpenAPI description. An org admin creates API tokens bound to a role and visibility scope. The API follows the same rules as the dashboards, so it never returns more than that role can see.
See the REST API reference for authentication, endpoints and examples.
Org hierarchy and director roll-ups
The Org Hierarchy report shows every department and sub-team with what each delivered. A node’s numbers include everyone placed under it, sub-teams too. That is the director roll-up: a director sees the whole department in one row.
An org admin builds the hierarchy under Admin → Hierarchy:
- Add nodes and place each person in one node. Placing one of a linked person’s identities places them all.
- Set a manager for a person to define the reporting line. Manager edges do not change any node’s numbers.
- Or import everything from a CSV with the columns
login,nodeand optionallymanager. A node is its path from the top, such asEngineering/Platform.
An import replaces the whole hierarchy. Every row is checked first, so one bad row changes nothing.
Open the result at Home → Org Hierarchy.
Attribute filters and group-by
Your admin can define contributor attributes such as employment type, level or location. Then:
- Open Efficiency → Attributes.
- Select Add filter → Group by and choose the attribute.
- Or use the Attribute filter to narrow a report to one value.
A slice with fewer people than the minimum slice size, 5 by default, is hidden from everyone. Slices span the whole workspace, so they do not combine with the team or group filter.