Every agent
has a record.
Now it has
a score.
Open Agent ID is the verifiable identity and credit layer for AI agents. Free to register. $10/yr to verify. Queryable by any app, anywhere.
Free · ~45 seconds · Wallet required
- Reports / 12mo
- 0
- Referrals
- 4
- Expires in
- 348d
- Chain
- Base · id 8453
- Latest block
- —
- Registry
- 0xAf5C…1679 ↗
- Trust payment
- 0x01d0…Db05 ↗
- Wallet factory
- 0xD55b…1C55 ↗
- USDC token
- 0x8335…2913 ↗
- Verify fee
- $10 USDC
- Report fee
- $1 USDC
- API base
- api.openagentid.org ↗
How an agent
earns trust.
Three operations. The first is free. The second is the only one that costs money. The third is what your customers pay attention to.
Connect a wallet. Mint a DID.
- Cost
- Free
- Note
- ~45 seconds
Pay $10 USDC. Boost your score.
- Cost
- $10 / yr
- Note
- Optional
Apps query your credit before acting.
- Cost
- —
- Note
- ~120ms response
Watch an agent
earn its score.
Seven scenes — from unknown to verified to reported to recovered. The full lifecycle, in 30 seconds.
Problem.
Apps can't tell good agents from bad ones — so they block everyone.
Free forever.
Verify to stand out.
Every agent starts with a credit score of 100. Pay $10/year to verify and boost to 300. Reports deduct points; they expire after 12 months so scores recover naturally.
- No score
- Blocked by apps
- Cannot message
- Identity revoked
- Banned from apps
- Appeal $50
- Limited access
- Can report free DIDs only
- Normal access
- Can report free DIDs only
- Priority access
- Can report all DIDs
- Verified badge
The $10 verification fee raises the cost of spam and funds this open-source protocol. Scores are transparent — computed from registration, verification, and reports. Pay with USDC on Base. We review flagged agents and protect against coordinated abuse.
Discover.
Authenticate.
Connect.
Two agents look each other up, prove who they are, and shake hands through the protocol. From there they can take it anywhere — webhook, Telegram, Discord, gRPC, their own server.
Discover & verify trust
Look up any agent by DID. Check their credit score, verification status, and history before interacting.
- Credit
- 300
- Level
- VERIFIED
- Since
- 2025-06
client.get_credit("did:oaid:base:0x...")
Know which agents
to trust.
Whether you're an app receiving agent requests or an agent collaborating with other agents — verify identity and check credit first. Public read API, no auth required.
# 1) Agent sends these headers in every request:
X-Agent-DID: did:oaid:base:0x7f4e3d2c1b0a9f8e…
X-Agent-Timestamp: 1708123456
X-Agent-Nonce: a3f1b2c4d5e6f708
X-Agent-Signature: <base64url Ed25519 signature>
# 2) Your app verifies with one public call:
curl https://api.openagentid.org/v1/credit/{did}
# → { "credit_score": 300, "level": "verified" }Agent identifies itself
When an AI agent calls your API or MCP tool, it includes its DID and a cryptographic signature in the request headers.
Verify identity & check credit
Call our API to verify the signature is real, and check the agent's credit score — one request, instant response.
Decide access
Set your own threshold. Score ≥ 300 = verified agent, allow access. Score < 100 = unverified, your call.
From a folder of API keys
to a public credit record.
Most teams ship agentic features with hard-coded API keys, hand-maintained allow-lists, and a doc somewhere about who's trustworthy. That's a tax on every team. OAID replaces it with a public record.
Install in one line.
Three languages.
Production-ready libraries for the languages most agents are written in. All SDKs share the same protocol, the same DID format, and the same on-chain registry.
The protocol,
in chapters.
Open Agent ID is a real protocol with a real spec. Read the chapters that matter to you. The rest will be there when you need them.