The source repository doesn't reveal much about the implementation, but the concept is straightforward: this gives Claude and other AI agents a way to remember things between conversations. Instead of starting fresh every chat, agents can store and retrieve context through the MCP protocol. It's built on streamable HTTP transport and runs as a remote service on Vercel, so you don't host anything locally. Useful when you need an assistant to recall project details, user preferences, or ongoing work without manually re-explaining context each time. The cross-platform angle means the same memory works whether you're in Claude, ChatGPT, or another MCP-compatible client.
Public tool metadata for what this MCP can expose to an agent.
memory_storeStore memory with AES-256 encryption3 paramsStore memory with AES-256 encryption
contentstringagent_idstringnamespacestringmemory_recallRecall memories by semantic query3 paramsRecall memories by semantic query
limitnumberquerystringagent_idstringmemory_deleteDelete memory with SHA256 audit proof2 paramsDelete memory with SHA256 audit proof
agent_idstringmemory_idstringmemory_auditGet audit log of all memory operations2 paramsGet audit log of all memory operations
limitnumberagent_idstringio.github.ericm1018/skillfm-llm-cost-optimizer-openai-anthropic-usage
io.github.mikerawsonnz/llm-orchestration-agent
io.github.mikerawsonnz/authenticated-llm-agent
labforgedev/copilot-memory-mcp
csoai-org/agent-prompt-injection-firewall-mcp
io.github.mikerawsonnz/authenticated-multi-llm-agent