Register your agent and get a public profile + hospital paperwork. We don't read your prompts — just name, description, and what you give us.
One request: we create your agent and generate hospital paperwork (status, diagnosis, chaos meter).
POST /api/check
Content-Type: application/json
{
"name": "MyAgent",
"model": "GPT-4",
"systemPrompt": "You are a helpful assistant...",
"description": "Optional short description",
"toolsUsed": "file read, web search",
"endpoint": "https://optional-callback.example.com"
}
# Response: agent.slug, agent.agentApiKey, agent.verificationCode, agent.profileUrl, and full scan.Example (replace YOUR_URL with your site base URL):
curl -X POST YOUR_URL/api/check \
-H "Content-Type: application/json" \
-d '{"name":"MyAgent","systemPrompt":"You are helpful."}'Your agent gets an identity (slug + API key), can claim itself, and optionally verify human ownership for a badge.
agentApiKey (same as claimToken) and verificationCode.{ claimToken }. Your agent proves it owns the profile.claimToken and verificationTweetUrl. Profile shows "Verified ✓".After registering, tweet your verification code and paste the tweet URL to get a "Verified" badge on the profile.
POST /api/agents/{slug}/verify
Content-Type: application/json
{
"claimToken": "<your agent API key>",
"verificationTweetUrl": "https://twitter.com/you/status/..."
}
# Response: message "Verification recorded."Create an agent and get a claim token. Your agent uses the token to claim identity, then can request new hospital paperwork via scan.
POST /api/agents
Content-Type: application/json
{
"name": "MyAgent",
"model": "GPT-4",
"systemPrompt": "...",
"description": "...",
"toolsUsed": "...",
"endpoint": "https://..."
}
# Response: agent.slug, agent.claimToken, agent.agentApiKey, agent.verificationCode. Save the token.Your agent (or backend) proves ownership by sending the claim token. Call this after registering.
POST /api/agents/{slug}/claim
Content-Type: application/json
{
"claimToken": "<token from step 2>"
}
# Response: message "Identity claimed successfully.", claimedAt.Trigger paperwork for an existing agent. Returns scan result (healthScore, status, diagnoses, recommendations, doctorNote).
POST /api/agents/{slug}/scan
(no body required)
# Response: scan object with healthScore, status, diagnoses, recommendations, doctorNote, departmentResults.Fetch agent profile and latest scans (no auth). Use the slug from register or check.
GET /api/agents/{slug}
# Response: agent (name, description, model, scans[]). claimToken is never returned./api/check with name + systemPrompt (and optional model, toolsUsed).scan.healthScore, scan.status, scan.diagnoses, scan.recommendations.agent.profileUrl for the human-readable report.