Skip to content
Sill Docs

Sill Docs

Sill is agent governance infrastructure. It answers three questions about every AI-agent interaction with a merchant site:

  • Identity — who is the visitor (signed agent card)
  • Intent — what they are trying to do (signed mandate)
  • Proof — what record remains (signed audit envelope)

These docs cover two product modes:

  • Discovery — a one-line embed script identifies AI-agent traffic, logs it to a signed audit envelope, and publishes per-site, signed, agent-readable manifests (A2A agent card, MCP server, ARD ai-catalog.json). Discovery is live in production.
  • Transactional — signed mandates, a policy engine, payment authorization through the merchant’s existing processor, and audit-grade logs. Validated end-to-end on the live Stripe rail in dogfood scope.

Sill never custodies funds. The merchant’s processor (e.g. Stripe) holds the card, authorizes the charge, settles, and pays out. Sill issues the signed authorization and the audit record.

Start here

Section titled “Start here”
  • What is Sill — the model, in one page.
  • Quickstart — install the Discovery embed (one script tag, paste before </body>).
  • Verify a signature — verify any Sill-signed manifest against the public JWKS, using only off-the-shelf ed25519 tooling.
Section titled “Related”