What Tablif read on Xonta's site
Xonta describes itself as: “Xonta Verifiable receipts of what your AI agent did”
- Price model: free tier
- Free tier: Free forever
- Stage: Live, no waitlist
- How it differs: “An agent's own log is a claim, not proof. Xonta makes the claim checkable.”
What it is
Calls itself “Trust layer for AI agents”.
- Calls itself: Trust layer for AI agents
“Xonta Trust layer for AI agents.”
Where it puts itself
Rules out “Nothing is sent to Xonta”.
- Says it will not: Nothing is sent to Xonta
“An agent's own log is a claim, not proof. Xonta makes the claim checkable.”
“Nothing is sent to Xonta — the gate, the keys and the log stay on your machine.”
What it costs
A free tier, free forever.
- Free plan: free forever
“Free and open source.”
Who it sells to
Sells to developers.
- Sells to: developers
“…node bin/xonta-gate.js -- npx -y @modelcontextprotocol/server-filesystem ~/Documents”
How it runs
A command-line tool. Runs locally and works offline. Can be self-hosted. Distributed through the npm. No API. Open source. AI is the core of the product.
- Kind: cli
- Runs where: on the device, offline · self-hosted
- Distributed via: npm
- API: no
- Open source: yes
- AI role: core to the product
“Xonta is a self-hosted gateway in front of any MCP server.”
“Nothing is sent to Xonta — the gate, the keys and the log stay on your machine.”
What it works with
Integrates with Claude Desktop and Cursor. Names MCP on its pages.
- Integrates with: Claude Desktop · Cursor
- Names: MCP
“…put Gate in front of a stdio MCP server (Claude Desktop, Cursor)”
“Xonta is a self-hosted gateway in front of any MCP server.”
What it offers as proof
Live, no waitlist. Claims “101/101 times a model claimed a check it never ran, under pressure”. Publishes docs.
- Number claimed: 101/101 times a model claimed a check it never ran, under pressure
“Run it yourself in one command.”
“101/101 times a model claimed a check it never ran, under pressure”
Sources
Tablif read these pages, last on 12 Nov 2026. Everything above is what the product says about itself; Tablif does not verify every claim.
Similar products in Tablif Town
- CoinGraph: Provable crypto data for AI agents, with a proof id
- Open Docy: The open-source Agent OS for decisions you can audit
- Zambo: The verifiable work layer for AI
- Open Agent Trust: Verifiable reputation for multi-agent systems
- Jid: The trust API for AI agents: check a business first
- Badge: Verified scores and certificates for AI agents
- dynos-work: Self-auditing AI coding agent that learns your codebase
- Slean: Verify AI research claims with certificates checked by the Lean kernel.