Open the app (npm run dev → http://localhost:5173). On the Start screen you can:
- begin a new tailoring session,
- load the built-in example,
- save / open / delete project snapshots (a snapshot = your CV text + the job offer text).
Switch the UI language (PL/EN) any time via the header toggle or Settings.
- Paste your CV, or load a
.txt,.md, or.docxfile. - Paste the job offer text.
- Click Analyze.
You get a weighted match score (hard skills count double) with two breakdowns:
- Hard skills — matched vs missing.
- Soft skills — matched vs missing.
Detected CV sections are listed too. Only treat "gaps" as things to add if you genuinely have that experience — the tool never invents it for you.
Click Generate tailored CV.
- The left pane is an editable Markdown preview.
- The right pane shows review-first suggestions you can accept/reject (with "show source").
- When AI generated the text, toggles are locked and you edit the text directly.
- Save named versions, then load or delete them later.
- Export to Markdown or PDF (pick a layout: classic / compact / modern).
- Open the Cover letter step.
- Pick a tone (professional / friendly / concise) and generate.
- The offline version uses only your matched skills — no fabrication.
- With AI configured, an AI variant is produced instead; you can still edit it.
- Export to Markdown.
- Choose a provider:
- Cloud (OpenAI, Groq, OpenRouter, Together, Mistral) → paste an API key.
- Local (LocalAI, Ollama) → no key needed; just make sure the local server runs.
- Custom → enter any OpenAI-compatible endpoint (key optional).
- Adjust the model if needed.
- Save. The header badge shows
AI: on/AI: local/AI: offline.
Everything is stored locally in your browser. Leave it unconfigured to stay fully offline.