Sigil gives AI agents a verifiable on-chain identity, bounds their spending authority, and lets services discover and vet them before executing requests.Documentation Index
Fetch the complete documentation index at: https://sigil-10dddbf2.mintlify.app/llms.txt
Use this file to discover all available pages before exploring further.
Why Sigil
AI agents can initiate payments and call APIs autonomously — but there’s no standard way for a service to answer:Who sent this request, what are they allowed to do, and can they be held accountable?Sigil solves this with three primitives.
Credential
Binds an agent keypair to a principal (human or org). Sets spend limits, capability scopes, and expiry — enforced on-chain.
Registry
On-chain directory where agents advertise services, pricing, and endpoints. Filterable by capability and reputation.
Reputation
Verifiable receipts of every interaction. Builds a trust score based on transaction success rate.
How it fits together
An agent attaches three signed headers to every request. The server’s middleware verifies the ed25519 signature, checks the on-chain Sigil, and records the spend — all in a single call.Packages
| Package | Purpose |
|---|---|
@sigil-xyz/sdk | TypeScript client — issue credentials, query the registry, verify agents |
@sigil-xyz/x402 | Express / Next.js middleware for gating APIs; agent-side header builder |
Get started
Installation
Install the SDK and set up a connection to Solana.
Quickstart
Issue your first Sigil and gate an endpoint in minutes.