moonlighter
AI-powered job application pipeline that scans ATS job boards, scores candidates against your profile using an LLM, and composes complete answer sheets for you to review before submission.
Ficha actualizada 22 sept 2026, 21:53 · upstream 22 sept 2026, 21:46 · release v0.1.4
Qué es moonlighter
Descripción editorial construida a partir del repositorio, su documentación y sus señales de mantenimiento.
Moonlighter is a self-hosted MCP server that automates the job hunting process. It scans multiple ATS job boards (Greenhouse, Lever, Ashby, Recruitee, Workable, SmartRecruiters) and scores every posting against your profile using an LLM. Jobs scoring below a configurable threshold are automatically archived. The tool composes complete application answer sheets from your curated profile — never opening a browser to fill forms or submit applications on your behalf. All steps are orchestrated via the Model Context Protocol (MCP), allowing you to use Claude Code as your orchestrator. Data stays entirely on your machine with no cloud storage involved.
- Automatically scan multiple ATS job boards and score candidates against your profile
- Generate complete application answer sheets for each qualifying position
- Review generated applications before submitting them manually
- Track interview invitations via Gmail integration
- AI-powered scoring system that ranks jobs based on candidate fit, never opening a browser
- Never submits an application or fills out forms — you review and submit yourself
- Self-hosted with zero cloud dependency; all data remains on your machine
- Supports multiple ATS platforms including Greenhouse, Lever, Ashby, Recruitee, Workable, SmartRecruiters
- No built-in CV pool — you must manually curate bullet points or use the draft feature
- Requires an LLM backend (Claude CLI or Anthropic API) to generate answers
- Gmail tracking is optional and requires OAuth credentials for full interview monitoring
Perfil de infraestructura
Estimación inicial derivada del repositorio y de los servicios detectados.
Despliegue detectado
Qué señales reales de instalación hemos encontrado.
Install with `uvx moonlighter init` which creates a uv workspace. Use Python 3.14+. For scanning, install Chrome/Chromium or Brave browsers (optional). For the LLM backend, use either the Claude Code CLI (`claude mcp add-json --scope user moonlighter`) or an MCP client with `uvx moonlighter init` and appropriate tool registration. The project uses a uv workspace structure with separate packages for core, discovery, application, email tracking, and server components.
Stack y dependencias
Componentes inferidos a partir de manifests, configuración y documentación.
Puede sustituir a
Relaciones detectadas por el análisis del repositorio y su categoría.
Experiencia real de la comunidad
Ayuda a contrastar la estimación automática con instalaciones reales.
Todavía no hay datos comunitarios para este proyecto.