Connect your AI agent to TestOrchestrator with MCP

02Connect a client in a minute

First, create a key: open Profile → API Keys, click Create API key, give it a label, and copy it. The key is shown exactly once and stored only as a hash — if you lose it, revoke it and make a new one. No admin needed.

Then point your client at the endpoint. Claude Code (CLI):

claude mcp add --transport http testorchestrator \
  https://api.testorchestrator.com/mcp \
  --header "Authorization: Bearer to_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"

Cursor (.cursor/mcp.json):

{
  "mcpServers": {
    "testorchestrator": {
      "url": "https://api.testorchestrator.com/mcp",
      "headers": {
        "Authorization": "Bearer to_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
      }
    }
  }
}

Claude Desktop and other clients that support remote MCP servers use the same shape with "type": "http". A client that can only launch local (stdio) servers can bridge to the endpoint with npx mcp-remote https://api.testorchestrator.com/mcp. Restart the client, then ask it something like "list the test cases in project 6" to confirm. The full walkthrough — every client and the security model — lives in the MCP setup guide.

03What your agent can do

The tools are grouped by area. We only advertise tools backed by a real capability, so everything your agent sees actually works.

• Discovery — resolve a project name to an id, then read its folders, fields, integration sources, cycles, environments, lifecycle states, statuses, tags, and users.
• Test cases & folders — list, search, create, and update cases (with steps and priority via custom field values), and build or reorganise the folder tree.
• Comments — add, read, edit, and delete comments on a case.
• Test runs — start and edit runs, record per-case results, read a run’s pass/fail/complete summary, and close it out.
• Cycles — create, update, and reopen the cycles (sprints/releases) that group those runs.
• Traceability — link a case to a Linear or Jira issue, see which cases reference an issue, and find what isn’t covered yet.

So a single agent can discover a project’s structure, author detailed test cases, comment, run and record a full test cycle, and trace coverage — without leaving the tool it already lives in. Two honest limits to set expectations: search is keyword/filter, not semantic, and the connection is POST-only (no server-push channel). CI-intelligence tools, test packs, and semantic search are not part of this release. The tool reference documents every tool and its arguments.

04A key that acts as you

The security model is the part we care about most. A key acts as its owner: it inherits exactly your access level, global role, project membership, and any per-project overrides. It can do only what you can do, in the projects you can access — never more, and never across workspaces. (This mirrors how Linear API keys inherit the issuer’s permissions.)

That makes granting an agent access a least-privilege decision by default: if you can’t edit a project, neither can your key. Keys are created and revoked by you, and revocation is immediate. The key value is never echoed back or written to logs, and resources outside your access return a flat "not found or access denied" so nothing leaks about other workspaces.

05Availability

The MCP server is rolling out. If you don’t yet see API Keys under your profile, it hasn’t reached your workspace — check back shortly, or read the setup guide so you’re ready to connect on day one.

Read the setup guide · Browse the tool reference · Start free