The agent layer for Radix

Paste any Radix address and learn its type (account, token, component, validator, package, NFT, …) plus core metadata. Components also reveal the blueprint they came from; pools show their pool unit and contents; internal vaults resolve to their owner. The place to start when all you have is an address.

Example What is component_rdx1cq… — a pool, an oracle, something else?

The full holdings of one account — fungible balances and the NFTs it owns. Empty balances are hidden by default. The everyday “what's in this wallet?” lookup.

Example What does account_rdx12… currently hold?

Name, symbol, icon, total supply and divisibility of a resource — plus its authorities (who may mint, burn, freeze or recall it, and the badge that gates each) and neutral factual flags (fixed supply, mintable, freezable…). Objective ledger facts, never a safety verdict.

Example Is this token's supply fixed, and who is allowed to mint it?

For one committed transaction: status, fee, timestamp and type; the net balance change per account; and the decompiled, human-readable manifest. Events and resource symbols are opt-in. Committed transactions never change, so results are cached permanently.

Example Who paid whom in transaction txid_…, and what fee did it cost?

The newest committed transactions, optionally scoped to one address and filtered exactly like the RadixScan dashboard: by type, accounts withdrawn-from / deposited-to, badges, resources, affected entities and a time window. Returns a compact summary per transaction; drill in with radix_get_transaction.

Example Show all transactions of this component in the last 24 hours.

Registration, effective fee, total XRD staked, uptime, website and the LSU / claim-NFT resources for one validator. Detailed mode adds the description, icon and current owner.

Example What fee and uptime does validator_rdx1s… have?

Every active validator on a network with rank, fee percentage and uptime. This is an address-less tool, so it takes an explicit network argument; the basis for deciding where to stake.

Example Which validators have the best uptime and a low fee?

Every function and method of a blueprint with its signature and access requirement — public versus role-protected, and which badge a protected call needs — plus the defined roles. Concise by default; ask for the full type trees when you need to build a typed call. Use it before building a manifest to know which calls a component supports and which the caller may actually invoke.

Example What methods can I call on this pool, and which need a badge?

A component's current stored state, decoded with field names — an oracle's prices, a dApp's configuration, an account's deposit rules. Works for components, pools, validators, accounts and access controllers, and reports who holds each role. Very large states are summarised to their top-level field names.

Example What prices are stored inside this oracle right now?

Reads a Scrypto KeyValueStore — the storage a component points at, e.g. an oracle's actual price entries. Page through all entries, or look up specific keys (address- or string-keyed). Values come decoded. The most mutable ledger state, so it is never cached.

Example List the individual price entries this oracle keeps.

The non-fungibles of a resource. Without ids: a page of the collection’s local ids and total count. With ids: the decoded data of those NFTs, and optionally who currently owns each. For one account’s NFTs use radix_get_account instead.

Example What traits does NFT #42 have, and who owns it now?

Every holder of a resource, sorted by holding descending, so the first page is the top holders. Works for fungibles (each holder's amount) and NFTs (each holder's count); reports the total and pages deeper. Backed by an optional gateway extension.

Example Who are the biggest holders of this token?

Given a badge resource (optionally one specific NFT id), every component, resource, pool or locker that names it as a role requirement — i.e. everything the badge holder can administer — each with its kind and, where available, blueprint and name.

Example Which components does this admin badge control?

Goal-oriented: describe the transfer and get a complete manifest (lock_fee, withdraw, deposit) for fungibles and/or NFTs. No manifest text to write by hand.

Example Build a manifest to send 100 XRD from account A to account B.

Takes a strict, typed array of steps and returns a canonical, validated manifest plus diagnostics. The backend generates the syntax, so an agent never hand-writes — and never mistypes — manifest text.

Example Withdraw 200 XRD, stake to this validator, deposit the stake units.

Runs a given manifest — even one you didn't build here — through the Radix Engine Toolkit for parsing and static checks. Read-only, accepts raw text: the sanity check before anything is signed.

Example Is this manifest someone sent me actually valid?

Simulates a transaction against the live Gateway with no signing and no cost: estimated fee, net balance changes and whether it would succeed, with a timestamp and validity window. So an agent never recommends a transaction blind.

Example If I run this swap, what would I receive and what is the fee?

The one prose tool: reads a manifest back in plain language, step by step — the basis for the confirmation a human sees before signing.

Example In plain English, what will this manifest do to my account?

The starting point for building transactions: how networks and addresses work and the full build → validate → preview → sign → submit lifecycle.

Example How does a Radix transaction get from draft to committed?

A catalogue of all manifest instructions, each with a short summary — the index for learning the manifest language.

Example What instructions can a manifest contain?

The full reference for a single instruction: its syntax, arguments and worked examples.

Example Show me exactly how TAKE_FROM_WORKTOP works.

A library of ready-made templates — transfer, stake, unstake, claim, mint, burn, swap — plus the {PLACEHOLDER} convention for filling them in.

Example Is there a ready-made template for unstaking?

The body of a single template with its placeholder slots — fill the slots with concrete values, then validate the result.

Example Give me the stake template so I can fill in amount and validator.

A directory of other agent-facing Radix services — e.g. DEX-aggregator MCP servers — for capabilities this server doesn't cover, such as resolving a token swap. Each entry lists the networks it supports. Informational only — a listing is not an endorsement.

Example Is there a service that can quote a token swap for me?

The inverse of resolve-address: for when you have a name but not an address. A small, curated register maps well-known Radix entities (tokens, dApp components) to their canonical on-ledger addresses, so an agent never has to guess one. Takes an explicit network and returns every matching candidate. Curated and informational — a result is the well-known address, not an endorsement or a safety verdict; confirm it on-ledger before signing.

Example What is the resource address of XRD on mainnet?

Returns a short URL; the sign page shows a plain-language summary plus a QR code / wallet deep link, and the human reviews and signs in their own Radix Wallet. Any account lock_fee is stripped automatically (the wallet adds its own fee). Keys never reach the agent. Stokenet always; mainnet when the deployment enables it.

Example Send this transaction to me to approve on my phone.

Poll a signature request by id: pending, committed (with the transaction intent hash), rejected or expired. On commit, the transaction is warmed into the cache so radix_get_transaction is immediate.

Example Did the human sign yet? If so, give me the transaction hash.