M
PRODUCT DETAILS
MELRA
Safer AI agent tool calls, with approvals and receipts
- Launch date · UTC
- 2026-10-19
- Domain
- —
- Domain ending
- —
- Category
- Artificial intelligence
About this product
When an AI agent writes to a file or calls an API, a timeout can leave it unclear whether the change happened. MELRA checks tool calls against local policy, pauses sensitive actions for approval, verifies the declared result, and keeps a redacted receipt. If a write's outcome is unknown, MELRA holds it for reconciliation instead of retrying blindly. Open-source alpha; requires Node.js 22.13+ or 23.4+.
Keywords
AI AgentapprovalOpen Source