MikeK184

Recollect — AI skill for Claude Code

AI community

Self-hosted AI memory and MCP coordination for coding agents.

How to install Recollect

This entry records only its repository, not the path inside it, so there is no exact command to give. Open MikeK184/Recollect and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Recollect does

Self-hosted AI memory and MCP coordination for coding agents. Persistent context across repositories and sessions, with knowledge graphs, hybrid search, and a desktop web UI. Built in Rust for Codex and Claude Code.

Alternatives in AI

  • Xerj — Local-first search engine and MCP server for AI agents 12.5k ★
  • Agentrq — AgentRQ: Human-in-loop realtime conversational task manager for AI Agents 1.1k ★
  • Codeseek — Rust-powered code intelligence CLI for AI coding agents 764 ★

README

Recollect

Memory and MCP tools for coding agents, with a desktop web UI.

Recollect keeps repository knowledge, documents, decisions and runbooks together so your next coding session can pick up where the last one left off. It runs on your own machine or a private server and connects to Codex and Claude Code.

What it does

  • Organizes knowledge into Brains, with access controls for individuals and teams.
  • Captures supported agent sessions and tool results, with sanitization and offline delivery.
  • Learns, revises and retires memories automatically under your chosen policies; manual review and corrections are optional.
  • Searches using exact matches, keywords, embeddings and graph relationships, with links back to the evidence.
  • Coordinates approved MCP tools locally, centrally or through private runners. Vault is an optional credential source.
  • Provides a desktop UI for exploring memory, graphs, processing and tool calls.

Built with Rust, React, PostgreSQL/pgvector and Neo4j/GDS Community.

Run locally

Install Rust/Cargo, Node.js/npm, Python 3.11+ and Docker with Compose, then run:

git clone https://github.com/MikeK184/Recollect.git
cd Recollect
./scripts/dev.sh

Open **http://127.0.0.1:8787**. The first run generates login credentials in the ignored `.env` file: use `RECOLLECT_OWNER_USERNAME` and `RECOLLECT_OWNER_PASSWORD`. The script builds the application and starts its databases, API and worker.

Learning and semantic search require an `OPENAI_API_KEY` in `.env` and an enabled Brain model policy. Model requests are billed by the provider. Each Brain starts with model transmission disabled.

For Docker installation and shared HTTPS access, see the [installation guide](docs/runbooks/installation.md).

Status and documentation

This is an early project for local and trusted private use. The first planned product scope is implemented and locally tested; desktop is the current UI target. The [evaluation report](docs/mappings/int