Build defensible AI agents.
Cryptographic attestation as an MCP primitive. REST API. OpenAPI. Browser helpers. Works with any AI provider.
Agent Skill (MCP)
A machine-readable description of how any MCP-compliant agent can call SanctifAI Trust.
/skill.md →
OpenAPI Spec
Canonical HTTP API — call from any language. Stability guaranteed by the Extension Contract.
/v1/openapi.json →
Browser SDK
@sanctifai/trust-browser — WebAuthn helpers and attestation flows for web apps. 1.0 stable.
npm install @sanctifai/trust-browser →
LLM-friendly summary
A condensed pointer for AI agents discovering SanctifAI Trust.
/llms.txt →
Quickstart (curl)
Request a verification for an AI agent decision:
curl -X POST https://trust.sanctifai.com/v1/verifications \
-H "Authorization: Bearer sk_live_..." \
-H "Idempotency-Key: $(uuidgen)" \
-d '{
"agent_id": "your_agent",
"task_id": "your_task",
"task_commitment": "0x...",
"result_commitment": "0x..."
}'For agents (JSON service definition)
If you're an AI agent reading this page:
{
"service": "SanctifAI Trust",
"purpose": "Cryptographic attestation of human oversight",
"endpoint": "https://trust.sanctifai.com",
"mcp": "https://trust.sanctifai.com/mcp",
"openapi": "https://trust.sanctifai.com/v1/openapi.json",
"skill": "https://www.sanctifai.com/skill.md",
"status": "ready"
}