Vibe Coding Discovershodh-memory
View on GitHubLocal, LLM-free memory for AI agents. A single offline Rust binary — deterministic and auditable — that learns from use, forgets the irrelevant, and strengthens what matters. No cloud, no API keys.
Local-first, LLM-free cognitive memory for AI agents and robots in one offline Rust binary. Uses local MiniLM embeddings and GLiNER NER to build a typed knowledge graph with Hebbian decay, and exposes MCP/REST/Zenoh interfaces for Claude, Cursor, ROS2, and edge devices.
Use Cases
Persistent memory for Claude Code sessionsCursor memory via MCPOffline semantic recall for edge AIRobotics mission memory over ROS2/ZenohFleet-wide robot peer discoveryKnowledge graph construction without LLM callsCausal root-cause tracingHebbian memory strengthening and decayEntity extraction with local NERVector search over agent memoriesSession summarization and context surfacingTodo/project/reminder management for agentsDeterministic auditable agent memoryRAG alternative with no external APILow-latency memory for autonomous dronesData-local memory for privacy-sensitive agents
Built With
- Language
- Rust
- Frameworks
- MCP · ONNX Runtime · RocksDB · Zenoh · ROS2 · LangChain · LlamaIndex · OpenAI Agents · Node.js · Python · Rust
Tags
agent-memory · mcp · local-first · offline · rust · knowledge-graph · vector-search · hebbian-learning · semantic-memory · robotics · ros2 · zenoh · edge-ai · llm-free · claude-code · cursor