mcp-memory-service and memora
These are competitors offering similar persistent memory solutions for AI agents via MCP, with the main differentiation being that the first provides a REST API for broader framework compatibility (LangGraph, CrewAI, AutoGen) while the second is MCP-native only.
About mcp-memory-service
doobidoo/mcp-memory-service
Open-source persistent memory for AI agent pipelines (LangGraph, CrewAI, AutoGen) and Claude. REST API + knowledge graph + autonomous consolidation.
Consolidates multi-agent memory using a knowledge graph with typed edges (causes, fixes, contradicts) and autonomous compression, accessible via REST API with ONNX-based embeddings that run locally. Implements Remote MCP support for browser-based claude.ai integration via Server-Sent Events, alongside traditional desktop MCP, with OAuth 2.0 authentication and self-hosted infrastructure (no cloud lock-in). Agent identity is tracked via `X-Agent-ID` headers for scoped retrieval, and conversation threading is preserved through `conversation_id` fields, enabling both shared memory across agent fleets and inter-agent messaging through semantic tag-based filtering.
About memora
agentic-box/memora
Give your AI agents persistent memory — MCP server for semantic storage, knowledge graphs, and cross-session context
Implements a Model Context Protocol (MCP) server with pluggable embedding backends (OpenAI, sentence-transformers, TF-IDF) and multi-tiered storage—local SQLite, Cloudflare D1, or S3/R2 with optional encryption and compression. Features include interactive knowledge graph visualization, RAG-powered chat with streaming LLM tool calling, event notifications for inter-agent communication, and automated memory deduplication via LLM comparison. Integrates with Claude Code and Codex CLI through stdio or HTTP transports.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work