# Mirix-AI/MIRIX

Mirix is a multi-agent personal assistant designed to track on-screen activities and answer user questions intelligently. By capturing real-time visual data and consolidating it into structured memories, Mirix transforms raw inputs into a rich knowledge base that adapts to your digital experiences.

Repository: https://github.com/Mirix-AI/MIRIX
Canonical: https://ross.abutalabs.com/products/mirix
Homepage: https://mirix.io/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: llm-agents, llm-memory, memory-agents, personal-assistant
Last push: 2026-08-20T13:59:50+00:00

## Health v2 (maintenance only)
Score: 72/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 60, longevity 36
- inputs: {"age_days": 509, "days_push": 13, "days_rel": 251, "gap_med": 15.5, "n_releases_24m": 5}
- flags: prerelease_only
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3434, forks 267 (observed 2026-08-28T04:08:04.304165+00:00)

## What it is
MIRIX is a self-hosted multi-agent personal assistant that captures on-screen activity and conversation to build structured long-term memory across six specialized memory types. It runs a Dockerized backend with a dashboard and API, and provides a Python client for interacting with the memory system.

## Use cases
- remember what I did on my computer last week
- build a personal knowledge base from screen activity
- add long-term memory to my LLM chatbot
- self-host an AI assistant with episodic and semantic memory
- search my past activity with full-text and vector search
- give my agent a memory system with core and procedural memory

## When to choose
- you want a self-hosted personal assistant that learns from screen observation
- you need a multi-agent memory architecture for LLM applications
- you want local, privacy-controlled long-term memory storage
- you want a Python client API to integrate memory into your own agents

## When to avoid
- you need a lightweight embedded memory library without a server
- you cannot run Docker or a PostgreSQL-backed backend
- you want a fully cloud-managed assistant with no setup
- you need non-Python client SDKs

## Facets
- artifact type: application
- maturity: active
- function: agent-framework, rag, llm-inference, search-engine, machine-learning
- domain: artificial-intelligence, large-language-models, developer-tools, self-hosted
- platform: self-hosted, python, cross-platform
- tags: llm-memory, memory-agents, personal-assistant, screen-tracking, multi-agent, llm-agents, ai-agents, docker, web-server

## Member repositories
- Mirix-AI/MIRIX (main) score 72

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:04.304165+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:37:43.272887+00:00, confidence not recorded.
  - readme: https://github.com/Mirix-AI/MIRIX (fetched 2026-08-28T04:08:04.304165+00:00, sha 3f1ca93c0e01)
  - homepage: https://mirix.io/ (fetched 2026-08-29T09:31:56.696549+00:00, sha 75d73288a056)
  - registry_pypi: https://pypi.org/pypi/mirix/json (fetched 2026-08-29T09:31:56.708228+00:00, sha cde03949d7dc)
- Data as of 2026-08-30T08:39:29.467469+00:00.