# Speakable
> The Trust Layer for AI Commerce. Speakable scores and certifies businesses for AI agent readiness — making products and services discoverable, verified, and safely transactable by every AI that shops.

## What Speakable Does
Speakable is an AI commerce readiness platform. It runs 15 specialized Claude agents that audit, score, remediate, and certify businesses across four pillars: Discoverability, Verification, Transactability, and Certification. The SpeakableScore (0-100) measures how ready a business is for autonomous AI agent commerce.

## Free AI Commerce Readiness Scan
Any business can get a free SpeakableScore. The Discovery Agent crawls your digital footprint and scores you across llms.txt, Schema.org markup, robots.txt AI crawler configuration, Answer Engine Optimization (AEO), Generative Engine Optimization (GEO), voice commerce readiness, and protocol compliance (UCP, Web Bot Auth, Visa TAP, Mastercard Agent Pay, Stripe ACP).

## Agent Infrastructure
- Agent Card: /.well-known/agent-card.json — A2A Protocol v0.3 compliant discovery endpoint
- Trust Certificate: /.well-known/speakable-trust.json — Ed25519-signed trust certification (includes WebMCP cert summary)
- UCP Profile: /.well-known/ucp — Universal Commerce Protocol v2026-04-08 capability declaration
- WebMCP Manifest: /.well-known/webmcp.json — W3C WebMCP browser-side tool surface (Speakable WebMCP Manifest v1.0, Ed25519-signed)
- Web Bot Auth Keys: /.well-known/http-message-signatures-directory — RFC 9421 signing keys

## API
- POST /api/v1/scan — Trigger a Discovery Agent scan (input: URL, output: SpeakableScore)
- GET /api/v1/scan/:scanId — Poll scan results
- GET /api/v1/verify/:businessId — Verify a business trust certification
- POST /api/v1/merchant/:clientId/negotiate — Initiate A2A negotiation with a merchant agent
- POST /api/v1/merchant/:clientId/intent — Submit an intent mandate for resolution

## The Four Pillars
1. Discoverability — Can AI agents find your products or services? Checks llms.txt, Schema.org, robots.txt, AEO, GEO, voice readiness, and WebMCP browser-side tool surface.
2. Verification — Do AI agents trust you enough to recommend you? Checks identity consistency, review authenticity, E-E-A-T signals, Web Bot Auth, and WebMCP manifest signing parity.
3. Transactability — Can AI agents safely purchase or book? Checks UCP compliance, payment protocol readiness (Visa TAP, Mastercard Agent Pay, Stripe ACP), mandate chain support, and WebMCP transactional tool surface.
4. Certification — Does AI have a reason to prefer you? Measures trust certification tier (Verified 60+, Trusted 75+, Certified 85+), proof density, verified review ratios, and signed WebMCP manifest with no runtime drift.

## Trust Certification Tiers
- Verified (60+): Automated scan + basic identity checks
- Trusted (75+): Verified + enhanced agent identity verification + review authenticity
- Certified (85+): Trusted + full payment protocol integration + continuous monitoring

## Company
- Name: Speakable (by Quicksilver AI)
- Website: https://getspeakable.ai
- Contact: support@getspeakable.ai
- Domain: getspeakable.ai

# Generated by Speakable (getspeakable.ai) — self-referential llms.txt