# Intuition-Lab/personal-model

Build your HUMAN.md.

Repository: https://github.com/Intuition-Lab/personal-model
Canonical: https://ross.abutalabs.com/products/personal-model
Homepage: https://intuition-lab.github.io/personal-model/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agent-memory, local-first, macos, mcp, personal-ai, personal-model, privacy, python
Last push: 2026-08-21T22:45:35+00:00

## Health v2 (maintenance only)
Score: 77/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 92, longevity 3
- inputs: {"age_days": 54, "days_push": 12, "days_rel": 52, "gap_med": 0, "n_releases_24m": 4}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1314, forks 95 (observed 2026-08-28T04:04:20.484019+00:00)

## What it is
Personal Model (Persome) is a local-first macOS runtime that captures focused cross-app activity and builds an evidence-linked personal memory model (HUMAN.md) of how you think and work. It exposes that model to trusted AI coding agents like Claude Code, Codex, and Cursor Agent via MCP, keeping all data private on your Mac.

## Use cases
- give my AI coding agent long-term memory about me
- build a HUMAN.md personal context file for Claude Code
- share one memory across multiple MCP clients like Codex and Cursor
- keep AI memory local and private on my Mac
- let my agent continue work with grounded, evidence-linked context
- inspect and correct what my AI knows about me
- import existing notes from Obsidian or Notion into agent memory

## When to choose
- you use macOS and want persistent, private memory for AI coding agents
- you want evidence-linked memory rather than opaque summaries
- you work across multiple MCP-compatible AI tools and want a shared context layer
- you value local-first data you can inspect, export, and delete

## When to avoid
- you are on Windows or Linux (macOS 13+ only)
- you need cloud-synced memory across devices
- you are uncomfortable granting macOS Accessibility and screen-capture permissions
- you want a turnkey hosted service rather than a self-managed local runtime

## Facets
- artifact type: application
- maturity: active
- function: mcp, agent-framework, ocr, nlp, privacy, self-hosted
- domain: artificial-intelligence, privacy, developer-tools, self-hosted
- platform: python, cli, self-hosted
- tags: agent-memory, personal-model, local-first, human-md, screen-context, mcp-server, claude-code, codex, cursor-agent, ai-agents, macos

## Member repositories
- Intuition-Lab/personal-model (main) score 77

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:20.484019+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 2026-08-30T04:48:44.602711+00:00, confidence not recorded.
  - readme: https://github.com/Intuition-Lab/personal-model (fetched 2026-08-28T04:04:20.484019+00:00, sha 32bfbce94503)
  - homepage: https://intuition-lab.github.io/personal-model/ (fetched 2026-08-29T12:07:19.384413+00:00, sha bba73397741a)
  - registry_pypi: https://pypi.org/pypi/personal-model/json (fetched 2026-08-29T12:07:19.394166+00:00, sha 382948b4d312)
- Data as of 2026-08-30T08:39:29.467469+00:00.