🎙️ OpenJujuSpeechKit

Free, open-source speech-to-text that runs on your machine. No API key, no usage limits, no audio ever leaves your computer.

Built and maintained by Juju Store → jujustore.app. When you need bigger models or cloud transcription, the same library upgrades seamlessly to the premium tier.

---

✨ What it does

	Free (local)	Premium (cloud via jujustore.app)
Transcribe audio → text	✅	✅
Runs offline on your machine	✅	—
Cost	Free, unlimited	Paid plans
Models	tiny · base · small · medium	pro · pro-large · turbo
Saved history & batch	—	✅
API key needed	❌	✅ (get one)

---

🚀 Install

# The library (lightweight):
pip install openjujuspeechkit

# ...plus the free local transcription engine:
pip install "openjujuspeechkit[local]"

The local engine (faster-whisper) downloads a small model the first time you run it. You also need ffmpeg installed (brew install ffmpeg on macOS, sudo apt install ffmpeg on Linux, or grab it from ffmpeg.org on Windows).

---

🏁 Quick start (free, local — no key)

In Python

from openjujuspeechkit import transcribe

result = transcribe("meeting.mp3")          # runs locally, free
print(result.text)

# Pick a bigger model for more accuracy:
result = transcribe("interview.wav", model="small", language="en")
for seg in result.segments:
    print(f"[{seg.start:.1f}s] {seg.text}")

From the terminal

openjuju transcribe meeting.mp3
openjuju transcribe interview.wav --model small --output transcript.txt
openjuju info        # shows models + links

That's the whole free tier. No sign-up, no key, no limits. 🎉

---

⭐ Premium (cloud) — when you need more

Get an API key at jujustore.app, then:

from openjujuspeechkit import Client

client = Client(api_key="juju_xxx")          # or set OPENJUJU_API_KEY
result = client.transcribe("longfile.mp3", model="pro")
print(result.text)

Premium runs on Juju Store's servers — bigger models, faster turnaround, and nothing to install or maintain.

---

📖 API reference

transcribe(audio_path, *, model="base", language=None, device="auto", compute_type="default")

Local, free transcription. Returns a TranscriptionResult with:

• .text — the full transcript (string)
• .language — detected language code
• .segments — list of Segment(start, end, text)

Free models: tiny, base, small, medium.

Client(api_key=None, *, api_url=..., timeout=300)

Premium cloud client. Reads OPENJUJU_API_KEY if api_key is omitted.

• .transcribe(audio_path, *, model="pro", language=None) → TranscriptionResult

Premium models: pro, pro-large, turbo.

---

🤝 Contributing

Issues and PRs welcome — see CONTRIBUTING.md. This project is MIT licensed and proudly open source.

💛 About

OpenJujuSpeechKit is a free gift from Juju Store — explore our other apps and the premium speech tier there.