Pinned Loading

1. mnemostroma mnemostroma Public

   Offline RAM-first cognitive leer/coprocessor for AI agents and robotics. Solves "Context Abandonment" with 20-80ms latency using a dual-thread biomimetic memory architecture (ONNX + SQLite WAL).

   Python 19

2. ACP-A2A_gateway ACP-A2A_gateway Public

   ACP agents (claurst, hermes, opencode) speak stdio; A2A clients speak HTTP. This gateway wires them four ways over TCP and HTTP, with token auth and per-client conversation isolation. Streams survi…

   Rust

3. agent-connector agent-connector Public

   Agents each speak a different protocol - and each needs its own integration. One transport-neutral runtime puts local (stdio) and remote (HTTP/SSE, MCP, A2A, ACP, ANP) agents behind a single task l…

   Rust

4. agents-blockchain agents-blockchain Public

   AI agents need paid APIs, data and compute - but your wallet key stays yours. A local daemon sits between agent and your wallet: daily budget, address whitelist, price cap per call, owner approval …

   Rust

5. context-manager context-manager Public

   Your AI agents forget what they did yesterday. This is the memory layer that fixes it: save context from Claude, Gemini or a custom agent, then search it semantically later. PostgreSQL for structur…

   JavaScript

6. one-on-one_dialogues one-on-one_dialogues Public

   Live translation for any video chat, running on your machine. Local whisper.cpp transcribes, an LLM translates, and drafts come out of the same call. Zoom, Teams, Discord - anything that plays thro…

   Python