KeyRunner Desktop
KeyRunner Desktop brings API development and production AI-agent workflows into one developer application. Build and verify an API request in API Workspace, export it as an AI tool, govern it with policies in AI Workspace, and inspect the agent executions that use it.
Choose a workspace
| Workspace | Use it to |
|---|---|
| API Workspace | Create requests, manage collections and environments, test responses, generate code, run flows, publish mock servers, and inspect history. |
| AI Workspace | Manage exported tools, compose policies, connect agents, inspect production activity, and explore execution traces. |
The project, workspace, and environment selectors in the header define the active development context. Universal Search (Ctrl+K) finds navigation destinations, collections, requests, tools, policies, agents, and workspaces.
Core developer workflow
- Build and test an API request in API Workspace.
- Replace environment-specific values with
{{variable}}references. - Select Export as AI Tool and describe the action and its inputs.
- Add the tool to a policy in AI Workspace.
- Assign that policy to an agent.
- Connect the agent through the KeyRunner SDK or MCP endpoint.
- Use Agent Observability to inspect model calls, governed tool calls, responses, latency, tokens, and cost.
Start here
- New to KeyRunner: Create your first request
- Moving existing API work: Import and export
- Building agent tools: Export an API request as an AI tool
- Operating an agent: Production agent workflow
- Diagnosing an execution: Agent observability and trace graph