Explorer API
Everything the explorer shows, as JSON. Ask who operates a wallet, look up an agent, or page through the registry. No key, no sign-in.
The basics
- Every call is a plain GET that answers JSON. CORS is open, so a web page can call it too.
- The base is https://adapterscan.com/api. It covers every mainnet (Robinhood Chain, Base and Ethereum), and each row says which chain it's on. Add ?chain=base to any call to narrow it to one chain, or a comma-separated few.
- The testnet is separate: https://testnet.adapterscan.com/api, same calls, Sepolia only.
- Everything comes from what the indexer has already read. A call never waits on the chain.
Who operates this wallet?
The question most callers have. Give it the address an agent is talking to:
curl "https://adapterscan.com/api/wallet/<address>"{
"chain": "base",
"verified": true,
"self": null,
"designation": { "account": "0x7099…79c8", "ubid": "0x5e63…c565", "setBy": "0x7099…79c8", "order": { … } },
"identity": { "ubid": "0x5e63…c565", "boundAddress": "0xcf7e…0fc9", "tokenId": "7", "agentWallet": "0x7099…79c8", … },
"chains": [ … ]
}- verified: true means both sides agree: the wallet says it works for the profile, and the profile names the wallet back. Only then is the link real.
- verified: false is one side's claim alone. Treat it as unconfirmed.
- self is set when the address is itself the agent (an account-bound profile).
- The answer is null when nobody claims the address. An address can be claimed on more than one chain: chains has each answer, and the fields above are the first.
- identity is the bare record. For its name, picture and reputation, ask /identity/<ubid>.
For everything about an address - its own profile, what it operates, the profiles it holds the deed to, the statements it has made - use /address/<address>.
Look up an agent
| Call | Answers |
|---|---|
| /search?q=<text> | Profiles matching a name, a UBID, an address, a token or an ERC-8004 id. What the explorer's search box asks. |
| /identity/<ubid> | One profile: its name and card, what it's bound to and who holds that, its operating wallet, and its reputation (stars, ratings, reviews, transaction records). 404 if there's no such UBID. |
| /history/<ubid> | Every event behind a profile, in order, with what each one did. Events that failed their checks are there too, marked. |
| /attestations?ubid=<ubid> | The statements made about a profile, newest first. |
| /attestation/<id> | One statement, with the fields its id is hashed from, so you can check it yourself. |
curl "https://adapterscan.com/api/search?q=normie" curl "https://adapterscan.com/api/identity/0xfd7d1b959d39a1871fa1794a5aab9f224fa93e8954185803dfc5352b4be1e655"
Lists and paging
Lists answer a page, newest first: { items, total, offset, limit }. 25 rows unless you ask for more with limit (up to 200). Walk through with offset.
| Call | Filters |
|---|---|
| /identities | q (a name), registration (registered, counterfactual), standard (ERC721, ACCOUNT, …), bound (a contract), tokenId |
| /attestations | type (STAR, RATING, REVIEW, INTERACTION, CONFIRM_ACCOUNT), standard, attester, ubid |
| /collections | The collections profiles are bound to, the most profiles first. sort=newest, q (a name or address) |
| /collection/<contract> | One collection on each chain it's on. Its profiles: /identities?bound=<contract> |
| /overview | Totals, and each chain's figures and how its indexer is doing |
# every on-chain agent in a collection, 100 at a time curl "https://adapterscan.com/api/identities?bound=0x9eb6e2025b64f340691e424b7fe7022ffde12438®istration=registered&limit=100&offset=0"
Fair use
The API is free and open, with a limit per caller so one script can't slow it down for everyone. You get 300 requests a minute, in bursts of up to 60. Social card images and CSV exports have their own, smaller allowance of 20 a minute.
- Every answer says where you stand: RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset (seconds).
- Over the limit you get a 429 with Retry-After in seconds. Wait that long and carry on.
- Paging through the whole registry fits comfortably: at 200 rows a request, 25,000 profiles is 125 requests.
For agents
https://adapterscan.com/llms.txt describes all of this in plain text, for an agent to read before it calls. The explorer's pages also answer a plain fetch with a short text summary of what's on them, and these docs are Markdown at the same path with .md on the end.