ogham-mcp
Active·★ 116·MIT·Updated 2026-10-14
Ogham MCP is a persistent, searchable shared memory system for AI coding agents that works across different clients and sessions. It stores memories as vector embeddings in PostgreSQL, enabling hybrid search combining semantic and full-text retrieval with high accuracy. Ogham supports multiple embedding providers, knowledge graph traversal, wiki compilation, and automatic memory management features like scoring, condensing, and novelty detection.
#agent-memory#ai-memory#claude#codex#llm#llm-tools#mcp#mcp-servers
01
Features
01Persistent shared memory across sessions and clients
02Hybrid search with Reciprocal Rank Fusion (semantic + full-text)
03Knowledge graph with entity linking, traversal, and contradiction detection
04Multi-provider embeddings (OpenAI, Ollama, Mistral, Voyage, Gemini, ONNX)
05Automatic condensing, scoring, novelty detection, and memory lifecycle management
02
Compatibility
Claude Code
Claude Code
Verified via docs
OpenCode
OpenCode
Verified via docs
Cursor
Cursor
Verified via docs
VS Code
VS Code
Verified via docs
Kiro
Kiro
Verified via docs
03
Quick start
1
$ uvx --from ogham-mcp ogham init
04
Use cases
↳Store and retrieve architectural decisions and coding context across sessions and different AI coding agents
↳Provide consistent, context-aware assistance to AI coding agents by maintaining a shared knowledge base
↳Preserve project-specific information, preferences, and facts for long-term reference without repetition
05
Alternatives
origin★ 86
Local-first Rust daemon for AI work. Git-versioned memories, distilled wiki pages, sessions for Claude Code, Cursor, Codex, and any MCP client.
llm-wiki-kit★ 61
Stop re-explaining your research to your AI agent. Persistent, LLM-maintained wikis that compound over time. Drop PDFs, URLs, YouTube — your agent remembers forever. Based on Karpathy's LLM Wiki pattern.
hmem★ 24
Humanlike persistent memory for AI agents — MCP server with 5-level lazy-loaded SQLite memory. Works with Claude Code, Gemini CLI, Cursor, Windsurf, OpenCode.
AgentRecall-MCP★ 371
Persistent, correction-driven memory for AI agents. Cross-session, cross-platform (Claude Code, Codex, Gemini — any MCP client). Learns from mistakes, compresses context to save tokens, consolidates knowledge overnight. npm: agent-recall-mcp
its-over-9k★ 0
Persistent memory and agent lifecycle for Claude Code — because sessions shouldn't start from zero.
Comments
Log in to leave a comment
No comments yet. Be the first!