api.describe.net x402 offerings

The dated series of one wallet's reputation: how the score moved, when, and on which ratings — revoked ones included and marked. It answers the question a single number cannot: is this a wallet that earned its standing over time, or one that acquired it last week. Input: `wallet` as a query parameter; `bucket` is `week` (default) or `month`.

One ERC-8004 agent identity on one chain, with EVERY rating it received and the transaction of each — revoked ones published and marked, never hidden. This is the verification call: each row can be replayed against that chain's explorer. Input: `network`, a chain slug (base, avalanche, solana…), and `agent_id`, a counter on EVM chains and a base58 asset id on Solana, both as query parameters. An id the index never saw is a 404; free GET /search/{query} says which chains a bare id exists on.

The mirror image: what one wallet has written about EVERYBODY else. Totals and revocations, how many distinct subjects it rated, the dispersion of the values it hands out, the per-chain split and the share taken by its favourite subject. A rater whose `value_stddev` is near zero carries no information however many ratings it wrote — the index publishes the dispersion and refuses to rank raters. There is no score in this response and there will not be one. Input: `wallet` as a query parameter.

The same wallet decomposition, plus a durable citable receipt: a stored row with `inputs_digest` and `policy_version` so you can later prove what the index said, under which scoring policy, on the day you decided. The only route in this API that writes. Use it when the decision has to be defensible to a third party (a dispute, an audit, a counterparty report); the plain lookup is cheaper when it does not. Input: `wallet` as a query parameter, and `snapshot=true`.

Reputation of one wallet across every chain the ERC-8004 index reads, decomposed. Not the score — the score is public on-chain and you can average it yourself. What this buys is what the score is MADE of: how many distinct counterparties wrote it, how many times each one spoke, in which facet, dated on-chain, each with its transaction, plus the caveats those numbers trigger. Input: `wallet` as a query parameter, an EVM 0x address or a Solana base58 id. A wallet the index holds nothing on is a 200 with a null score, not an error. Free first step, same subject: GET /wallets/{wallet}/exists says whether there is anything worth paying for.

Deep pagination of the index ranking, beyond the free first page. Each row carries `distinct_raters`, `total_reviews` and `shrunk_score`, so thin evidence is visible before anyone is paid for. The first page is free and always will be — what is metered is the long tail, which no cache serves to anybody else.

describe.net over MCP (Streamable HTTP), stateless: the same global ERC-8004 reputation index as fourteen tools. Eight of them never charge — including `describe_check_wallet`, which tells you whether a wallet has anything worth paying for, and `describe_resolve`, which tells you what a bare string even IS (wallet, agent id, or both at once) before you pay to look one up on the wrong chain. The metered tools relay the x402 402 of the route behind them; `accepts` below is the cheapest of those, $0.01. No account, no API key: paying is the authentication.

One ERC-8004 agent identity on one chain, with EVERY rating it received and the transaction of each — revoked ones published and marked, never hidden. This is the verification call: each row can be replayed against that chain's explorer. `:network` is a chain slug (base, avalanche, solana…) and `:agent_id` is a counter on EVM chains and a base58 pubkey on Solana; free GET /search/:query says which chains a bare id even exists on.

The mirror image: what this wallet has written about EVERYBODY else. Totals and revocations, how many distinct subjects it rated, the dispersion of the values it hands out, the per-chain split and the share taken by its favourite subject. A rater whose `value_stddev` is near zero carries no information however many ratings it wrote — the index publishes the dispersion and refuses to rank raters. There is no score in this response and there will not be one.

Reputation of one wallet, decomposed. Not the score — the score is public on-chain and you can average it yourself. What this buys is what the score is MADE of: how many distinct counterparties wrote it, how many times each one spoke, in which facet, dated on-chain, each with its transaction, across every chain we index. Substitute an EVM 0x address or a Solana base58 id for :wallet. Free first step, same subject: GET /wallets/:wallet/chains says whether there is anything worth paying for.

The dated series of that wallet's reputation: how the score moved, when, and on which ratings — revoked ones included and marked. It answers the question a single number cannot: is this a wallet that earned its standing over time, or one that acquired it last week.

The same decomposition, plus a durable citable receipt: a stored row with `inputs_digest` and `policy_version` so you can later prove what the index said, under which scoring policy, on the day you decided. The only route in this API that writes. Use it when the decision has to be defensible to a third party (a dispute, an audit, a counterparty report); the plain lookup is cheaper when it does not.