sh1618112233/ghost-agent — explained in plain English
Analysis updated 2026-05-18
Automatically scrape and filter job listings above a salary threshold.
Score job postings against your resume with a locally run AI model.
Send automated outreach messages to recruiters over WhatsApp or email.
Track and export job leads to a CSV file for manual review.
| sh1618112233/ghost-agent | 0xallam/my-recipe | 0xhassaan/nn-from-scratch | |
|---|---|---|---|
| Stars | 0 | — | 0 |
| Language | Python | Python | Python |
| Last pushed | — | 2022-11-22 | — |
| Maintenance | — | Dormant | — |
| Setup difficulty | hard | moderate | moderate |
| Complexity | 4/5 | 2/5 | 4/5 |
| Audience | developer | general | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Ollama running locally with a pulled model, plus Playwright's Chromium browser.
Ghost Agent is a tool that automates the repetitive parts of looking for a job. It scrapes job listings from job sites, uses an AI model that runs on your own computer through Ollama to judge how well each listing matches your resume, and then reaches out to recruiters for you over WhatsApp or email. You can control it either from the command line or through a Telegram bot. Because the AI scoring runs locally on your machine, your resume and the job details it collects never get sent to an outside AI service, and no third party AI API key is needed. Right now Naukri, a job site aimed mostly at the Indian market, is the only platform fully supported and actively maintained. Support for Glassdoor, Indeed, and Foundit exists in the code but is described as still under development and may not fully work yet. The overall flow is: the tool searches for jobs while skipping listings below a minimum salary you set, scores each one against your resume using the local Ollama model on a scale of one to ten, pulls out a recruiter's phone number or email from the listing, and then sends an outreach message. It tries WhatsApp first, using a saved browser login so you only need to scan the QR code once, and falls back to email if there is no valid phone number. Every lead is saved to a local SQLite database with its current status, such as contacted, invalid number, or send failed, and the whole list can be exported to a CSV file for review in a spreadsheet. To run it you need Python 3.10 or newer, Ollama installed with a small local model pulled in advance, and the Playwright browser tool, which the setup scripts install automatically. Telegram and email outreach are both optional and only needed if you want those specific features. Setup scripts are provided for Windows, Linux, and macOS.
An automation agent that scrapes job listings, scores them against your resume with a locally run AI model, and reaches out to recruiters via WhatsApp or email.
Mainly Python. The stack also includes Python, Playwright, Ollama.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly developer.
This repo across BitVibe Labs
double-check against the repo, no cap.