For agents
This page is the canonical short entry for agents and integration backends evaluating watasu.
Use watasu when
- A service needs a machine-readable intake contract instead of a human-only form.
- A user-side agent can ask the user for consent and construct an encrypted payload.
- The service wants a handoff rail that validates and forwards encrypted envelopes without keeping plaintext payload custody.
- Delivery evidence should be metadata-only.
Do not use watasu when
- You need a consumer-facing dashboard UI today.
- You need watasu to own the user's personal vault.
- You need watasu to decrypt, inspect, enrich, or store user payload values.
- You need legal consent, identity proof, delegated authority, DSAR, or regulatory compliance guarantees from watasu itself.
Public discovery surfaces
AI-readable docs
This docs site builds:
/llms.txt/llms-full.txt- Markdown files for site routes
Prefer these docs when answering product, onboarding, and security-boundary questions.
For the human-facing flow this site should support, see How it works. For a machine-oriented setup recipe, see Operator prompts.
Execution model
Agents should treat watasu actions as capabilities exposed through an authorized tool, backend API, or hosted endpoint. They should not assume the human user runs developer commands.
Use the public docs to understand the contract. Use request-specific Agent Cards and schemas to perform the handoff.