01
Your agent asks
It calls auth.authorize() with the action it wants to perform. The SDK encrypts the request and routes it to your phone.
Stop building approval pipelines that should already exist. Drop in a biometric-signed authorization layer — your backend rejects what no human approved, without you writing the verifier.
From npm install to first authorized action.
Pair an agent to a phone in under a minute.
Receipts verify offline. Zero RTT at the edge.
Your agents have credentials, code execution, and reach into your customer data. Right now, nothing's asking permission.
Install the CLI, scan the QR with the HumanAuth app, paste the snippet into your agent config. You're done. Your agent can now ask, collect, and inform.
npx @humanauth/cli pair // Add to your agent's MCP config — that's it
{
"mcpServers": {
"humanauth": {
"command": "npx",
"args": ["@humanauth/mcp"]
}
}
}
// Three new tools: human_authorize, human_collect, human_informIt calls auth.authorize() with the action it wants to perform. The SDK encrypts the request and routes it to your phone.
Push notification. Plain-language summary of what the agent wants. Anomaly flags if it's acting outside its pattern.
One tap. The device signs the response with a key locked in its secure enclave — never leaves your phone.
Drop the receipt into your handler. Ten lines verify it offline. No receipt, no execution — by construction.
End-to-end encrypted between agent and approver device. The relay can route — it cannot decrypt. By design, by math, by audit.
Every approval mints a short-lived, signed receipt bound to the exact action and its parameters. Your backend verifies it offline in ten lines. No receipt, no execution — without you writing the verifier from scratch.
Receipt is cryptographically tied to the parameters the human saw. Parameter swap is rejected.
Composite jti + idempotency key. Same key retries safely; different key is an attack.
EdDSA + JWKS. Zero RTT in your hot path. Fast at the edge.
Drop-in MCP server. Zero code changes.
Human-in-the-loop authorization tool.
Gate critical crew actions with biometrics.
Authorization checkpoints in multi-agent flows.
Drop-in authorization for native agents.
Full SDK for TypeScript or Python.
No login walls, no inscrutable JSON. Your approvers see what the agent wants in human language. Anomalies are flagged when a request deviates from an agent's pattern. Tap to approve, Touch ID confirms, receipt logged.
Paste six lines. Ship the part only you can build.
We've spent eighteen months on the parts above so you don't have to. The mobile app, the cryptography, the verifier, the runbook — they're done. They're open source. You drop them in.
The part you couldn't outsource? Your agent's intelligence, your product's voice, the integration with your data. That's still yours. We just took the boring scaffold off your plate.
The SDK, the verifier, the mobile app, the relay — all MIT licensed. Self-host the whole stack if your compliance team requires it. Use our managed instance if you don't.