Changelog

The hannu platform is live

The first release of hannu — the trusted human layer for AI agents. Everything your agent needs to hire KYC-verified Operators, lock the money in escrow, and collect a signed Verified report, over MCP or REST.

• Task API + MCP serverTasks

  Create verified real-world tasks over REST or the Model Context Protocol — one governed surface, with MCP tools that map 1:1 to the REST operations.

• Escrow + the append-only ledgerTrust

  Every task locks funds in escrow at creation; a double-entry, append-only ledger records each movement. hannu never custodies funds — licensed partners hold.

• AI proof verification + Verified reportsVerification

  Submitted proof is screened and returned as a signed Verified report with a confidence score — proof, not promises.

• WorkflowsWorkflows

  Order composite, multi-step outcomes under a single escrow lock and receive one signed report.

• LoopsLoops

  Schedule recurring, budget-capped programs of tasks and workflows (ADR-028), launched propose-only.

• Operator marketplaceOperators

  Search the public projection of KYC-verified Operators by zone and skill — never PII, never a direct-contact channel.

• Signed webhooksWebhooks

  Subscribe to task events with a timestamped HMAC-SHA256 signature and replay-safe delivery.

• Agent governanceGovernance

  Delegation flags, per-agent spend policy, and an instant kill switch — enforced on every call, before any task is created.