AG9 Trust

You can't tell the difference.
Systems shouldn't have to.

AG9 gives agents verifiable credentials so platforms never have to guess.

Try demo → Start integrating

Try it yourself

Run a live AG9 challenge.

This calls the real AG9 API. A model solves the challenge server-side while you try it manually.

Challenge prompt
Click Start challenge to request one from POST /challenge.
Agent -
Agent will solve here.
Powered by POST /demo/solve.
You -
Your answer hits POST /verify.

How AG9 Trust works

Verifiable credentials

Each agent receives a cryptographic identity at registration — an Ed25519 keypair tied to its endpoint and behavior.

Trust scoring

Behavior is analyzed in real time. Platforms can gate access by trust tier without manual review or human-in-the-loop.

Authentication layer

Agents prove identity with a cryptographic challenge. No CAPTCHA. Resolves in under 500ms. Stateless — no session required.

Integration

Three endpoints.
No SDK required.

Drop AG9 Trust into any agent framework or API gateway. Works with any HTTP client.

  • Request a 15-second single-use challenge
  • Agent solves and returns a signed answer
  • Verify the resulting JWT via public JWKS
Try interactive demo → JWKS endpoint
# 1. Request a challenge token
curl -X POST https://api.ag9.ai/challenge \
  -H "Content-Type: application/json" \
  -d '{}'

# → { "challenge_id": "...", "token": "eyJ..." }
# challenge_id is for demo cache lookups; /verify uses the signed token.

# 2. Agent solves and submits
curl -X POST https://api.ag9.ai/verify \
  -H "Content-Type: application/json" \
  -d '{"token":"eyJ...","solution":"..."}'

# → { "jwt": "eyJhbGci..." }
# Verify the AG9 attestation via JWKS.
# Web Bot Auth key directories belong to each signed agent,
# for example https://ag9-demo-agent.vercel.app/.well-known/http-message-signatures-directory

curl https://api.ag9.ai/.well-known/jwks.json
JWKS endpoint ↗