Try it on your own task.

See whether explicit evidence checks improve your workflow before buying. This is a real local tool loop—not a scripted success demo.

Requires Node.js 22+ and an MCP-capable host. The official Community repository includes the pinned release and setup instructions.

Install the free Community editionJoin the five-user proof pilot

No purchase is required. Pilot feedback is opt-in through a public GitHub issue; never include private task content or credentials.

One task. One checkable result.

1. Pick a small task whose result you can inspect: checking one document, testing a function, or reviewing a local file. Give it an exact target name.

2. Ask your agent to start a scoped Living Stack session, authorize the bounded action, do the work with its existing tools, and record the observed outcome with a verification reference.

3. Ask it to check its completion claim against that same target and the returned outcome ID. Then try the wrong target: the successful outcome must not support an unrelated claim.

The intended contrast is a supported claim versus an unrelated one being rejected. The result must come from your actual run. A receipt reference alone does not independently establish that an event happened.

Want to resume the whole workflow tomorrow?

Community keeps the proof ledger, but does not provide searchable context, saved workflow checkpoints, portable signed exports, release-byte checks, or Team coordination.

Complete Local adds those tools. Save the evidence and next step before stopping, restart your host, then load that same session's checkpoint and revalidate its context.

Get Complete Local · $19 onceSee the paid workflows

If the seven-tool loop covers your needs, keep using it free. There is no expiry-based upsell and no automatic charge.