easy-transcriber-stt

Local-first audio transcription tool. Upload audio or video, get a transcript in seconds. No cloud required if you use local providers.

---

Features

• Multi-provider transcription — OpenAI (whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe), ElevenLabs (scribe_v2)
• Local providers (optional) — faster-whisper, Qwen3-ASR, Ollama — no API key required
• Audio Lab — advanced preprocessing: noise reduction, vocal isolation (Demucs), HQ pipeline
• Dual denoise engine — DeepFilterNet (AI-based) or ffmpeg afftdn (fast, no GPU required)
• Real-time progress — WebSocket-powered stepper with ETA
• Transcription history — stored locally in SQLite
• Multiple output formats — TXT, SRT, VTT, JSON
• Bilingual UI — English and Italian
• No CDN, no Node.js — all static assets vendored, works fully offline
• Double-click startup — start.bat (Windows) or start.sh (macOS/Linux)

---

Requirements

• Python 3.10+
• ffmpeg installed and on PATH
• At least one provider API key or a local provider installed (see docs/providers.md)

---

Quick Start

git clone https://github.com/agiuseppe28/easy-transcriber-stt.git
cd easy-transcriber-stt
python start.py

Or double-click start.bat (Windows) / start.sh (macOS/Linux).

The app opens automatically in your browser. On first launch, a setup wizard will guide you through provider configuration.

---

Providers

Provider	Models	Requires API Key	Notes
OpenAI	whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe	Yes	Best accuracy
ElevenLabs	scribe_v2, scribe_v1	Yes	Speaker diarization
faster-whisper	tiny → large-v3	No	Local, GPU optional
Qwen3-ASR	0.6B, 1.7B	No	Local, lightweight
Ollama	any speech model	No	Local, requires Ollama

Full setup instructions: docs/providers.md

---

Configuration

Copy .env.example to .env and add your API keys:

cp .env.example .env
# Edit .env with your keys

OPENAI_API_KEY=sk-...
ELEVENLABS_API_KEY=sk_...

Settings (language, default provider, output formats) are managed through the UI and saved to settings.json.

---

Documentation

• Installation guide — detailed setup for all platforms
• Providers — compare providers, get API keys, configure local models
• FAQ — common questions

---

Contributing

See CONTRIBUTING.md. Issues and PRs are welcome.

---

License

MIT — see LICENSE.

---

---

Italiano

Strumento di trascrizione audio locale. Carica audio o video, ottieni una trascrizione in pochi secondi. Non richiede cloud se usi i provider locali.

Caratteristiche

• Trascrizione multi-provider — OpenAI (whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe), ElevenLabs (scribe_v2)
• Provider locali (opzionali) — faster-whisper, Qwen3-ASR, Ollama — senza chiave API
• Audio Lab — preprocessing avanzato: riduzione rumore, isolamento voce (Demucs), pipeline HQ
• Progresso in tempo reale — stepper WebSocket con ETA
• Storico trascrizioni — salvato localmente in SQLite
• Formati di output — TXT, SRT, VTT, JSON
• Interfaccia bilingue — italiano e inglese
• Nessun CDN, nessun Node.js — funziona completamente offline
• Avvio con doppio click — start.bat (Windows) / start.sh (macOS/Linux)

Requisiti

• Python 3.10+
• ffmpeg installato e nel PATH
• Almeno una chiave API provider oppure un provider locale installato

Avvio rapido

git clone https://github.com/agiuseppe28/easy-transcriber-stt.git
cd easy-transcriber-stt
python start.py

Oppure doppio click su start.bat (Windows) / start.sh (macOS/Linux).

L'app si apre automaticamente nel browser. Al primo avvio, un wizard guida la configurazione del provider.

Per la documentazione completa vedi la sezione inglese sopra.