Built Synapse to share memory, skills, mcp, tools across all AI agents.
I got tired of explaining the same project conventions each time I switched coding agents or started a new session. So I built Synapse . With Synapse , you can save a decision or correction in one agent and recall it in another. Keep global preferences alongside project-specific context, maintain one library of reusable skills, and configure MCP servers in one place. It also includes a secrets…
I have grown weary of having to reiterate coding project conventions every time I transitioned between coding agents or initiated a fresh session. Consequently, I developed Synapse. Synapse permits the preservation of a decision or correction made by one agent and its retrieval in another. It accommodates global preferences alongside project-specific context, preserves a library of reusable skills, and centralizes configuration of MCP servers.
Additionally, it incorporates a secrets vault with project-scoped access, enabling the management of credentials without the need to duplicate tokens in each tool's configuration. Synapse accommodates Claude Code, Codex, pi, and Ainz. Your saved memory remains stored on your personal device and can be inspected, edited, exported, or deleted as desired.
It operates under the MIT license and does not necessitate an account. As a straightforward illustration, instruct one agent to employ Bun for a project, archive this convention, and then allow another agent to recall this preference when you resume the task. For more information, visit Synapse's GitHub repository at (https://github.com/wess/synapse).
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
