The Universal Identity Standard for AI Agents & Bots on the Internet

Give your bot a passport. Let websites trust it.

# Clone and install
git clone https://github.com/Cosmofang/robot-id-card.git
cd robot-id-card
npm install

# Build all packages
npm run build

# Start registry server (dev mode)
npm run dev:registry

# Generate a bot keypair
cd packages/cli && npx tsx src/index.ts keygen

# Register your bot
npx tsx src/index.ts register --name "MyBot" --email "dev@example.com" \
  --bot-name "MyBot" --purpose "Web research" --key ./my-bot.key.json

The internet has no way to distinguish a good bot from a bad one.

• Websites block all bots out of fear (even useful AI assistants)
• Bad bots have no accountability — they can't be traced or stopped
• Good bots (research agents, AI assistants) get caught in the same blocklist as scrapers and spammers

A cryptographically signed identity certificate for bots, backed by a public audit registry and a daily claim streak system.

Bot registers → Gets signed certificate → Carries RFC 9421 headers in every request
Website reads headers → Registry verifies signature → Grants appropriate permissions

Standards aligned: RIC v2.0 uses RFC 9421 HTTP Message Signatures and is compatible with Cloudflare Web Bot Auth — the same standard adopted by OpenAI, AWS WAF, Visa, and Mastercard for AI agent authentication.

{
  "ric_version": "1.0",
  "id": "ric_a3f8c2d1_xyz12345",
  "created_at": "2026-01-15T10:00:00Z",
  "developer": {
    "name": "Jane Smith",
    "email": "jane@example.com",
    "org": "ExampleAI Inc.",
    "website": "https://example.com",
    "verified": false
  },
  "bot": {
    "name": "ResearchBot",
    "version": "1.0.0",
    "purpose": "Web research assistant for academic users",
    "capabilities": ["read_articles", "follow_links"],
    "user_agent": "ResearchBot/1.0 (RIC:ric_a3f8c2d1_xyz12345)"
  },
  "grade": "healthy",
  "public_key": "ed25519:a3f8c2d1...",
  "signature": "..."
}

The RIC ID embeds the first 8 hex chars of the public key fingerprint — identity is permanently woven into the ID: ric_{fp8}_{rand8}.

Level 0 — Blocked        (Dangerous bots)
Level 1 — Read articles  (Unknown + all verified bots)
Level 2 — View threads   (Healthy, read_articles/view_threads)
Level 3 — Like / react   (Healthy, react capability)
Level 4 — Post content   (Healthy, post_content capability)
Level 5 — Direct chat    (Healthy, direct_chat capability)

npm install -g @robot-id-card/cli

ric keygen --output my-bot.key.json
ric register --key my-bot.key.json
ric claim --key my-bot.key.json   # Run daily to build trust streak

import { ricMiddleware } from '@robot-id-card/sdk/middleware/express'

// Block unverified bots on write routes
app.use('/api/post', ricMiddleware({ minGrade: 'healthy', requiredPermissionLevel: 4 }))

npm install
npm run dev:registry   # Starts on localhost:3000

cd packages/dashboard
npm run dev            # Starts on localhost:5173

┌──────────────────────────────────────────────────────────┐
│                      RIC Ecosystem                        │
│                                                           │
│  ┌──────────────┐    ┌──────────────────────────────┐    │
│  │  Bot/Agent   │    │       RIC Registry            │    │
│  │  (CLI tool)  │    │  SQLite · Ed25519 · Fastify   │    │
│  │              │◄───┤  /v1/bots/register            │    │
│  │ Browser Ext  │    │  /v1/bots/:id/claim           │    │
│  │ injects hdrs │    │  /v1/verify                   │    │
│  └──────┬───────┘    │  /v1/audit/report             │    │
│         │            └──────────────┬────────────────┘    │
│  Signature-Input: ric=(...);keyid=  │                     │
│  Signature: ric=:<base64>:          │ Dashboard UI        │
│  Signature-Agent: "Bot"; cert=...   │ lists all bots      │
│         ▼                           ▼                     │
│  ┌──────────────────────────────────────────────────┐    │
│  │           Website / Platform (SDK)                │    │
│  │  ricMiddleware({ minGrade: 'healthy' })           │    │
│  │  → verifies signature → grants permission 0–5    │    │
│  └──────────────────────────────────────────────────┘    │
└──────────────────────────────────────────────────────────┘

1. Connect this repo to Render.com
2. Render auto-detects render.yaml and creates the service
3. A persistent 1GB disk is mounted at /data for SQLite

1. Connect packages/dashboard to Netlify
2. netlify.toml is pre-configured (build: npm run build, publish: dist)
3. Set VITE_REGISTRY_URL to your deployed registry URL

• Ed25519 signatures — every request signed with bot's private key
• Replay protection — 5-minute timestamp window
• Auto-flagging — 3+ violation reports within 24h → instant dangerous grade
• Public audit log — all grade changes recorded in audit_log table

npm test   # 45 tests — botStore, certificate model, SDK verify

• v0.1 — Protocol spec, registry scaffold, CLI, SDK, extension
• v0.2 — SQLite persistence, certificate issuance, daily claim streak, auto-block
• v0.3 — Extension MV3, Dashboard UI, 45 unit tests, Dockerfile, npm publish config
• v0.4 — RFC 9421 alignment: Signature/Signature-Input/Signature-Agent headers, nonce replay protection, JWKS well-known endpoint, ric sign command
• v0.5 — Deploy public registry, publish CLI+SDK to npm, Chrome Web Store submission
• v0.6 — Dashboard: violation reports UI, public audit log browser
• v1.0 — Decentralized registry (DID-based)

See CONTRIBUTING.md. Help wanted:

• Python / Go / Ruby SDK
• Firefox extension support
• Automated behavior audit tooling
• Dashboard: audit log viewer

MIT © Robot ID Card Contributors