Runs tool-using AI agents on your own hardware with shell, file, search, and web tools, judge-graded approvals, and a cluster dashboard, using the models you choose.
Use it now
Keep it with up to 9 other tools you're setting up, then share them all as one link.
Try this first
Ask your coding agent:
- Built-in Agent Tools
Use Turnstone to run shell commands: ‹what you need›
- Intent Validation
Use Turnstone to grade every tool call: ‹what you need›
- Parallel Workstreams
Use Turnstone to run parallel agent workstreams across cluster nodes: ‹what you need›
Install the CLI
pip install turnstone
Setup guide: github.com/turnstonelabs/turnstone/blob/main/QUICKSTART.md
Call the API
Auth A bearer token from Turnstone Labs
Setup guide: github.com/turnstonelabs/turnstone/blob/main/QUICKSTART.md
Compare
Press Compare to add it to the tray at the bottom of the window, then pick up to 2 others anywhere on the site.
- Access
- CLIAPI
- Auth
- Bearer Token
- Skills
- 5
Skills
Built-in Agent Tools
Runs shell commands, reads and edits files, searches, and fetches web pages as model tools, with approval before each call runs.
Intent Validation
Grades every tool call with an LLM judge running your model, giving a risk assessment and evidence before the call runs.
Parallel Workstreams
Runs parallel agent workstreams across cluster nodes, routed by the console's rendezvous proxy and shown on a live dashboard.
Connect MCP Servers
Connects external MCP tool servers with native deferred tool loading, or BM25 tool search as a fallback for other models.
Discord and Slack Channels
Routes conversations from Discord and Slack to agent workstreams through the Turnstone channel gateway and its adapters.
