#254254th of 489 in Skills, plugins and prompts

memora

MCP memory server with graph relations and semantic search for coding agents

Stars
749
License
MIT
Last commit
Oct 2026
Last release
Sep 2026
Language
Python
Stars, last 29 weeks+357
25 Apr 2026: 392 stars2 May 2026: 394 stars9 May 2026: 396 stars16 May 2026: 401 stars23 May 2026: 401 stars30 May 2026: 402 stars6 Jun 2026: 406 stars13 Jun 2026: 407 stars20 Jun 2026: 412 stars27 Jun 2026: 416 stars4 Jul 2026: 419 stars11 Jul 2026: 427 stars18 Jul 2026: 430 stars25 Jul 2026: 430 stars1 Aug 2026: 433 stars8 Aug 2026: 585 stars15 Aug 2026: 655 stars22 Aug 2026: 688 stars29 Aug 2026: 700 stars5 Sept 2026: 709 stars12 Sept 2026: 717 stars19 Sept 2026: 723 stars26 Sept 2026: 725 stars3 Oct 2026: 729 stars10 Oct 2026: 733 stars17 Oct 2026: 742 stars24 Oct 2026: 744 stars31 Oct 2026: 746 stars7 Nov 2026: 747 stars14 Nov 2026: 749 stars

392 on 25 Apr 2026749 on 14 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Memora is an MCP server that stores agent memories in SQLite, with optional sync to S3/R2 or Cloudflare D1. It supports semantic search (TF-IDF, sentence-transformers, OpenAI embeddings), typed graph edges, and an absorb step where an LLM classifies new facts as duplicate, update, contradiction, related or new. A built-in graph viewer and a topic digest tool are included.

Who it is for: Developers giving Claude Code or Codex persistent cross-session memory

Strengths

  • Updates supersede old facts with lineage kept, rather than overwriting or deleting them
  • Local SQLite via pip install, or D1 and S3/R2 storage with optional encryption
  • One HTTP process can serve multiple stores at /mcp/<name>
  • Tool profiles (full 44, leader 19, agent 12) limit which tools clients see

Weaknesses

  • Container path requires Apple's container CLI on Apple silicon with macOS 26
  • Container setup is manual: credential file, proxy script and LaunchAgent
  • Default embeddings and absorb use OpenAI; OpenRouter without an embedding pair falls back to TF-IDF
  • Local embeddings need the [local] extra, about 2 GB for PyTorch

What it needs

  • GPU optional
  • Docker
  • Needs SQLite, Cloudflare D1 (optional), S3/R2 (optional), OpenAI API (default embeddings and LLM), Apple container CLI (container deployment)
  • Models: OpenAI embeddings, sentence-transformers, TF-IDF
  • port 8765

What changed

From the release notes. No flag means the note does not say so, not that the upgrade is safe.

Also in Skills, plugins and prompts

See all 489
Also in Skills, plugins and prompts
RankProjectScore
1claude-memPersistent cross-session memory plugin for Claude Code and other coding agents118.1k starsApache-2.097 out of 100
2impeccableDesign skill with 24 slash commands for AI coding agents' frontend output Live demo ↗ (opens in a new tab)112.4k starsApache-2.097 out of 100
3gstackMarkdown slash-command skills that give Claude Code planning, review, QA and ship roles141.9k starsMIT96 out of 100
#4rufloAgent harness adding swarms, memory and hooks to Claude Code and Codex Live demo ↗ (opens in a new tab)82k starsMIT96 out of 100
#5eccAgents, skills, hooks and memory bundle for Claude Code and other harnesses305.7k starsMIT95 out of 100