JobFit AI is a Gradio app that reads a candidate CV, runs a Kimi 2.6 agent through the OpenAI Agents SDK, searches the web with Olostep, and produces a ranked Markdown job-fit report.
The app is aimed at AI, data science, technical writing, technical content, curriculum, and developer education roles. The preferences box can be changed for other job searches.
- Reads a CV PDF with
pypdf. - Runs a Kimi 2.6 agent through the OpenAI Agents SDK.
- Gives the agent two tools:
search_jobsandread_job_page. - Uses Olostep Search to find job listings.
- Uses Olostep Scrape to read selected job pages.
- Streams simple progress logs for tool calls, parameters, and tool output sizes.
- Generates a Markdown report with a best match, ranked jobs, job notes, and rejected jobs.
Install dependencies:
pip install -r requirements.txtSet the required API keys.
PowerShell:
$env:MOONSHOT_API_KEY="your_moonshot_key"
$env:OLOSTEP_API_KEY="your_olostep_key"macOS/Linux:
export MOONSHOT_API_KEY="your_moonshot_key"
export OLOSTEP_API_KEY="your_olostep_key"python app.pyOpen:
http://127.0.0.1:7860
Upload a CV PDF, edit the job preferences if needed, then click Generate JobFit Report.
- Read the uploaded CV.
- Build one prompt from the CV and job preferences.
- Run the
JobFit AIagent with Kimi 2.6. - The agent searches for job listings with
search_jobs. - The agent reads selected listings with
read_job_page. - The agent writes the final Markdown report.
- The app displays the report in Gradio.
Progress appears in the log box while the app runs. The generate button is disabled during a run to prevent duplicate requests.