ToolMesh

•
6
•

Governs and audits AI agent interactions with enterprise systems by providing a secure self-hosted gateway.

ToolMesh acts as a crucial control layer, transforming potentially uncontrolled AI agent tool calls into a secure, governed, and auditable process. It enables AI agents to safely interact with real-world enterprise systems by offering features like fine-grained authorization, secure credential injection, and comprehensive audit trails. Furthermore, its unique DADL (Declarative API Description Language) allows developers to describe any REST API as an MCP tool using simple YAML, effectively replacing the need for custom MCP wrapper servers and accelerating integration.

주요 기능

01Describe any REST API as an MCP tool using DADL YAML files.
02Enforce fine-grained authorization with OpenFGA for user-to-tool access control.
03Implement input and output gating using JavaScript policies to manage data flow.
04Securely inject credentials at execution time, never exposing them in prompts.
05Maintain a comprehensive and queryable audit trail for every tool call.
062 GitHub stars

사용 사례

01Securely connect AI agents to internal enterprise systems and APIs.
02Centralize governance and auditing for all AI agent tool calls.
03Efficiently integrate diverse REST APIs and multiple MCP servers for AI agent access.