Fill job applications with a local model and a final human review
At a glance
Lazy Apply fills repeat application fields from a CV, asks about sensitive answers, and leaves the final submission to the applicant. This workflow is listed under job-applications, forms, local-ai.
- Agents / tools
- Lazy Apply, Chrome, Gemma 3, Ollama
SourceBuilding a tool for these workflows? See weekly rates.
What they did
1) Install Chrome 116+, Node 20+, pnpm, and Ollama. Clone https://github.com/adybose/job-applier and enter its folder. Pull `gemma3:4b-it-qat` with Ollama, then run `pnpm install` and `pnpm build`. 2) Allow extension requests in Ollama: on macOS run `launchctl setenv OLLAMA_ORIGINS "chrome-extension://*"` and restart Ollama. 3) In chrome://extensions enable Developer mode and Load unpacked → extension/output/chrome-mv3. Complete its onboarding checks. 4) Import your CV PDF or LinkedIn PDF, review the extracted profile, and resolve conflicts. 5) Open an application, open Lazy Apply, press Start, and grant access to that job site. Review salary, visa, consent, and drafted-answer cards; handle uploads, login, or CAPTCHA handoffs yourself. 6) Inspect the table of changed fields and submit the application yourself.
How it went
The author reports roughly five minutes of personal effort per application, versus 15–20 minutes previously.
Worth knowing
The extension never submits. The setup above uses the macOS Ollama-origin command; other systems need their corresponding environment setting.
