The agent layer for Radix

The clearest, safest way for AI agents to read, understand, and build transactions on the Radix network.

Radix transaction manifests are human- and machine-readable, statically analyzable, and free to simulate. An agent can explain exactly what a transaction will do before a human signs it.

llms.txt llms-full.txt

Why Radix for agents

Readable transactions

Manifests are typed, bash-like instructions — an agent can parse, check and explain intent directly.

Static analysis

Worktop simulation and type checks catch mistakes before a human is ever asked to sign.

Free preview

Every transaction can be simulated against the Gateway — fees, balance changes, success — so an agent never acts blind.

Native resources

Tokens are first-class on the ledger, not bespoke contracts — fewer edge cases, less to hallucinate.

See it

This is a complete “stake 200 XRD” transaction. No decoding required — the agent (and the signer) can read exactly what happens:

CALL_METHOD
  Address("account_rdx12…")
  "lock_fee"
  Decimal("0.5")
;
CALL_METHOD
  Address("account_rdx12…")
  "withdraw"
  Address("resource_rdx1tk…xrd")
  Decimal("200")
;
TAKE_FROM_WORKTOP
  Address("resource_rdx1tk…xrd")
  Decimal("200")
  Bucket("xrd")
;
CALL_METHOD
  Address("validator_rdx1s…")
  "stake"
  Bucket("xrd")
;
ASSERT_WORKTOP_CONTAINS
  Address("resource_rdx1t…lsu")
  Decimal("195")
;
TAKE_ALL_FROM_WORKTOP
  Address("resource_rdx1t…lsu")
  Bucket("stake_units")
;
CALL_METHOD
  Address("account_rdx12…")
  "deposit"
  Bucket("stake_units")
;

Tools implemented

The Read+Build MCP server exposes focused, network-scoped tools over the Model Context Protocol. Building is keyless; signing stays human-in-the-loop. The public endpoint is live.

Read

  • radix_resolve_address
  • radix_get_account
  • radix_get_resource
  • radix_get_transaction
  • radix_get_validator
  • radix_get_validator_list

Build & simulate

  • radix_build_transfer
  • radix_build_manifest
  • radix_validate_manifest
  • radix_preview_transaction
  • radix_explain_manifest

Manifest docs

  • radix_manifest_workflow
  • radix_manifest_instructions
  • radix_manifest_instruction
  • radix_manifest_templates
  • radix_manifest_template

Ecosystem

  • radix_list_ecosystem_services

Signing human-in-the-loop

  • radix_request_signature
  • radix_signature_status

Endpoint: mcp.ai.radixscan.io — connect over Streamable HTTP at /mcp or SSE at /sse

Principles

  • Keyless read & build — the agent never holds private keys
  • Signing is human-in-the-loop, non-custodial (Radix Wallet)
  • Free preview before any recommendation
  • Every tool is explicitly scoped to mainnet or stokenet
  • No personal data, no user tracking — only anonymous, aggregate operational metrics (no IP, user, or address)

Privacy Notice| Terms and Conditions