DEDUPED Hot AI repos. Real use cases. What people build. Get the Monday 5

How to use MemPalace: what you can do with it

What is MemPalace? MemPalace is a local-first AI memory system. It stores your conversation history as verbatim text and retrieves it with semantic search, without summarizing or paraphrasing. People and projects become wings, topics become rooms and original content lives in drawers, so searches can be scoped. ChromaDB is the default backend, the retrieval layer is pluggable, and it ships a CLI, an MCP server, a Docker image and skills for coding agents. Nothing leaves your machine unless you opt in.

How do I use it? Run uv tool install mempalace to put the MemPalace CLI in an isolated environment on your PATH (or pipx install mempalace); then run mempalace init ~/projects/myapp with the path to your own project; run mempalace mine ~/projects/myapp to mine the project files into the palace. Full steps below.

5 people shared what they built with MemPalace. See what people are building

Trended earlier

View repo on GitHub

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer

What do you want to do?

3 use cases

  • Give Claude Code searchable memory of your past sessions

    Install the MemPalace CLI, mine your Claude Code transcripts and project files into a local palace, then search them and load context into a new session.

    View prompt
    Install MemPalace/mempalace with `uv tool install mempalace`, run `mempalace init` on this project, then mine my Claude Code sessions with `mempalace mine ~/.claude/projects/ --mode convos`. Before answering any question about past decisions in this repo, run `mempalace search` with the question and quote the matching original text. Do not delete or move any transcript files.

  • Run a Docker MCP server for Claude Code memory

    Pull the multi-arch MemPalace image and wire it into an MCP client as a stdio server, with the palace persisted in a Docker volume.

    View prompt
    Set up MemPalace/mempalace as a stdio MCP server for Claude Code using the Docker image: run `docker pull ghcr.io/mempalace/mempalace:latest`, then add a mempalace entry to my MCP config that runs the image with -i, the mempalace-data volume on /data and my Claude Code projects folder mounted read-only at /transcripts, using an absolute path. Show me the config before saving it.

  • Let your coding agent install agent memory through skills

    Add the MemPalace skills, then have your coding agent detect your system, install the package, configure MCP and verify the connection.

    View prompt
    I ran `npx skills add MemPalace/mempalace`. Use the mempalace skill to set up MemPalace as a private local palace on this machine: detect my system, install the package with uv tool if available, configure MCP and verify the live connection. Tell me every system change before you make it, and leave weekly release checks disabled.

What people are building with MemPalace

  • Built withthe web

    Added persistent memory to a Claude Code Telegram bot with the MemPalace MCP server

    He wired the MemPalace MCP server into his fork of claude-code-telegram with a 7 line MCP config and a CLAUDE.md telling Claude to search and save memories. After a full /new session reset the bot still recalled his name and projects, with zero code changes to the bot.

    MemPalace on GitHubTry it: Give Claude Code searchable memory of your past sessions

    By Ivan Morgillo (ivanmorgillo.com) on the web

    View original post on the web

  • BenchmarkGitHub

    Reproduced the MemPalace LongMemEval benchmark on an M2 Ultra and wired it into an orchestrator

    On a Mac Studio M2 Ultra the raw mode hit 0.966 R@5 in 4:37, matching the README, while aaak and rooms modes scored 0.842 and 0.894. He then mined 905 MB of Claude Code history into a 76,839 drawer palace for his Sandcastle orchestrator over MCP.

    MemPalace on GitHub

    By @gizmax on GitHub

    View original post on GitHub

  • Built withDEV

    Built hermes-mempalace, a local memory provider plugin for Hermes Agent

    He wrote a plugin that plugs MemPalace into the Hermes Agent memory provider lifecycle, buffering turns, mining them into a palace per Hermes profile and giving the agent mempalace_search, mempalace_wake_up and mempalace_status tools, with no API key or extra server.

    MemPalace on GitHub

    By Abdallah Deeb (@abdallah) on DEV

    View original post on DEV

  • Built withDEV

    Built an OpenCode plugin that auto-saves every conversation to MemPalace

    He replaced a fragile cron sync with opencode-mempalace-persistence, a TypeScript plugin that uses the chat.message and session.idle hooks to save each turn to MemPalace asynchronously, with mempalace-mcp registered as a local MCP server in OpenCode.

    MemPalace on GitHub

    By geco (@gecojs) on DEV

    View original post on DEV

  • Hands-onthe web

    Ran a three layer Claude Code memory stack with CLAUDE.md, MemPalace hooks and its knowledge graph

    Across four active projects he pairs CLAUDE.md with MemPalace stop and precompact hooks that file session memories into project wings, and adds 3 to 5 knowledge graph triples at the end of each session, invalidating old facts so the graph keeps a timeline.

    MemPalace on GitHubTry it: Give Claude Code searchable memory of your past sessions

    By Guilherme Rondon (guirondon.dev) on the web

    View original post on the web

Real builds shared on X, Hacker News, Reddit, DEV and GitHub, summed up by Deduped from each post and its links; each card links to the original post. See what people are building with other repos

Quick start with MemPalace

  1. Check you have what it needs: Python 3.10 or newer; uv or pipx recommended for the CLI (plain pip only inside a virtualenv); A vector-store backend (ChromaDB by default, no configuration).
  2. Open MemPalace/mempalace on GitHub and read the README and license.
  3. Run uv tool install mempalace to put the MemPalace CLI in an isolated environment on your PATH (or pipx install mempalace)
  4. Run mempalace init ~/projects/myapp with the path to your own project
  5. Run mempalace mine ~/projects/myapp to mine the project files into the palace
  6. Run mempalace mine ~/.claude/projects/ --mode convos to mine your Claude Code sessions (scope with --wing per project)

These steps come from our top use case, Give Claude Code searchable memory of your past sessions, which also has a copy-paste agent prompt.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for MemPalace/mempalace

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Best AI repos for Software and DevOps

About this repo

  • Who it is forDevelopers using Claude Code, Codex, Cursor, Gemini CLI or other MCP clients who want their agent to remember past sessions and project decisions without sending data to a cloud memory service.
  • Why it is hot60k stars, talked about in 2 AI sources we track
  • LanguagePython
  • Sourcegithub.com/MemPalace/mempalace ยท Back to Hot repos

Requirements

  • Python 3.10 or newer
  • uv or pipx recommended for the CLI (plain pip only inside a virtualenv)
  • A vector-store backend (ChromaDB by default, no configuration)
  • About 300 MB of disk for the embedding model
  • Docker only if you use the container image

Add this badge to your README

Maintain MemPalace? Show readers where to find these use cases. Use cases on Deduped

Markdown 

[![Use cases on Deduped](https://dedupedai.com/badge/MemPalace/mempalace.svg)](https://dedupedai.com/use-cases/mempalace-mempalace/?utm_source=github&utm_medium=badge&utm_campaign=MemPalace-mempalace)

HTML 

<a href="https://dedupedai.com/use-cases/mempalace-mempalace/?utm_source=github&amp;utm_medium=badge&amp;utm_campaign=MemPalace-mempalace"><img src="https://dedupedai.com/badge/MemPalace/mempalace.svg" alt="Use cases on Deduped" height="20"></a>

A static image, no tracking script. About the badge

Questions about MemPalace

What is MemPalace?

MemPalace is a local-first AI memory system. It stores your conversation history as verbatim text and retrieves it with semantic search, without summarizing or paraphrasing. People and projects become wings, topics become rooms and original content lives in drawers, so searches can be scoped. ChromaDB is the default backend, the retrieval layer is pluggable, and it ships a CLI, an MCP server, a Docker image and skills for coding agents. Nothing leaves your machine unless you opt in. It is an open-source project on GitHub: https://github.com/MemPalace/mempalace.

How do I use MemPalace?

Run uv tool install mempalace to put the MemPalace CLI in an isolated environment on your PATH (or pipx install mempalace); then run mempalace init ~/projects/myapp with the path to your own project; run mempalace mine ~/projects/myapp to mine the project files into the palace. The use cases on this page give the full steps and a copy-paste agent prompt for each job.

Who is MemPalace for?

Developers using Claude Code, Codex, Cursor, Gemini CLI or other MCP clients who want their agent to remember past sessions and project decisions without sending data to a cloud memory service.

What can I do with MemPalace?

3 use cases so far, for example: Give Claude Code searchable memory of your past sessions; Run a Docker MCP server for Claude Code memory; Let your coding agent install agent memory through skills. Each has steps and a copy-paste agent prompt.

What do I need to run MemPalace?

Python 3.10 or newer; uv or pipx recommended for the CLI (plain pip only inside a virtualenv); A vector-store backend (ChromaDB by default, no configuration); About 300 MB of disk for the embedding model.

What are alternatives to MemPalace?

Other hot AI repos in Software and DevOps: Deep Infer, Harness2, Orca, Opencode, Graphify.

Is MemPalace safe to run?

Deduped does not own, audit or vouch for it. Read the README and license, run it on your own systems first and review every command before you run it.