Van-to-Quote

Описание и применение

Van‑to‑Quote – автономное приложение, которое преобразует устные заметки о работе в готовый PDF‑квот, используя Whisper и локальную модель Qwen2.5, без передачи данных в облако.

Задача

Tradespeople spend evenings typing quotes after jobs; cloud quoting apps require internet and store customer data.

Результат

The app enables offline, private, zero‑per‑quote cost quoting, reducing manual typing and eliminating cloud dependence.

Рабочий процесс

• 1. Record spoken job notes using the app’s microphone or text box.
• 2. Whisper (faster-whisper) transcribes the audio to text.
• 3. A local LLM (Qwen2.5 3B via Ollama) parses the transcript and outputs JSON with items, quantities, and hours.
• 4. Python code matches JSON items against a local prices.json file, applies VAT, and calculates totals.
• 5. ReportLab generates a PDF quote; items not in the price list are flagged for manual pricing.
• 6. The user reviews and edits the draft in a Flask web page, then downloads the final PDF.

Доказательства

Van-to-Quote is a local app that turns a tradesperson's spoken job notes into a priced PDF quote.

Источник

Whisper (faster-whisper) transcribes -> a local LLM extracts items, quantities and hours as JSON -> your own price list supplies every price and all maths is plain code.

Источник

The key design choice: the model never sets a price. It only decides what was done.

Источник

I tested the full flow end to end on my own laptop with spoken job descriptions.

Источник