Resume Matcher

Resume Matcher is an open-source project started by Saurabh Rai. The README calls it a self-hosted AI harness that works with Claude, ChatGPT, DeepSeek, Kimi, GLM, Gemma and other local or remote models. You upload your master resume, paste a job description, and it rewrites the resume for that job and produces a cover letter as a PDF. It can also generate email templates and interview preparation material. The README says the rewrite must not add skills, tools, employers or metrics that are missing from your master resume. You can queue several job postings in the background, and the queue picks up again after a restart. AI Inspiration World found no feature that applies or sends email for you; you copy the email template and send it yourself.

Who it's for:Job seekers who want a tailored resume for each posting and prefer to keep their data on their own computer. You install it yourself with the command line or Docker, which will be hard going if you have not done that before.

Before you start: Run it with Docker, or install the Python tool uv and Node, then start it with uv sync, uv run app and npm install. Windows has its own setup guide. The interface is in English, Spanish, Chinese, Japanese, Brazilian Portuguese, French and Korean. Resumes and settings are stored in a local SQLite database. If you choose a remote provider such as OpenAI, Anthropic, Gemini, DeepSeek, OpenRouter or Groq, the relevant resume and job data go to that provider; with a local Ollama model the work stays on your computer. By default the service listens only on 127.0.0.1 and has no login. SETUP warns that anyone who can reach it can read your resumes and change your AI settings, so keep it off the public internet. SETUP also offers installing uv by piping a remote script from curl straight into sh; AI Inspiration World does not recommend that, and the same document gives pip install uv as an alternative. The software is free and open source, takes donations and needs no account. Remote models are billed by their providers, and local models use your own hardware. Apache-2.0 license.

Based on the project's public documentation. Not tested by AI Inspiration World.