Skip to content

OpenWebUI_README.md

Latest commit

 

History

History
101 lines (69 loc) · 3.46 KB

OpenWebUI_README.md

File metadata and controls

101 lines (69 loc) · 3.46 KB

🚀 Quick Start: Local Audio Transcription with LIRA & OpenWebUI on NPU

Ready to turn your browser into a voice-powered AI playground? With LIRA and OpenWebUI, you’ll be transcribing audio locally in no time—no cloud required! This guide shows you how to connect everything for instant, private voice transcription right from your desktop.

💡 Tip: Make sure you’ve checked out the main README.md for setup and installation of Ryzen AI + LIRA.

1. Set up environments

Recommended: Use separate conda environments to avoid dependency conflicts.

  • For LIRA, reuse ryzen-ai-1.5.0 to leverage NPU support.
  • For OpenWebUI, create a new environment.

LIRA setup:

Follow the instructions in the Getting Started section of the main README.md to install and set up the Ryzen AI environment.

conda activate ryzen-ai-1.5.0
lira serve --help

OpenWebUI setup:

In a new environment, let's set up OpenWebUI:

conda create -n openwebui python=3.11 -y
conda activate openwebui
pip install open-webui

2️. Start OpenWebUI 🌐

open-webui serve

OpenWebUI will be available at http://localhost:8080 by default.

3️. Start the LIRA Whisper server 🗣️

Activate the LIRA environment and start the server using the CLI tool. The server exposes an OpenAI-compatible transcription route at /v1/audio/transcriptions.

To target Whisper on CPU:

lira serve --backend openai --model whisper-base --device cpu --host 0.0.0.0 --port 5000

For NPU:

lira serve --backend openai --model whisper-base --device npu --host 0.0.0.0 --port 5000

Once running, your API endpoint will be:

http://localhost:5000/api/v0

4️. Configure OpenWebUI for LLM and Audio Transcription 🛠️

  1. Open your browser and navigate to http://localhost:8080
  2. Sign-in with your credentials

To enable chat and audio transcription features in OpenWebUI, connect it to your local LLM server (for chat) and LIRA server (for speech-to-text).

a. Connect OpenWebUI to your LLM server

  • Go to ⚙️ Admin Settings → Connections → OpenAI API
  • Click ➕ Add Connection - URL: Enter your LLM server endpoint (e.g., for Lemonade: http://localhost:8000/api/v0) - API Key: Leave blank or use a placeholder like fake-key if required.
  • Click Save

b. Connect OpenWebUI to your LIRA ASR server

  • Go to ⚙️ Admin Settings → Audio
  • Under Speech-to-Text, select OpenAI - URL: Enter your LIRA server endpoint (e.g., http://localhost:5000/v1)
  • Save your settings

🎉 You can now chat with your local LLM and transcribe audio using LIRA directly in OpenWebUI.
Record from your mic or upload audio files (.wav, .mp3)—OpenWebUI will send them to LIRA for transcription.

📝 Notes & Tips

  • If you exported a Whisper ONNX model to a custom directory, set LIRA_MODEL_DIR before starting the server, or use lira serve flags to point at the export.
  • For NPU runs, start lira serve from ryzen-ai-1.5.0 so Vitis AI tooling and drivers are available.
  • If running behind a reverse proxy, update OpenWebUI's API Base URL accordingly.

See the main README.md for full LIRA setup and model export instructions.

Copyright (C) 2025, Advanced Micro Devices, Inc. All rights reserved.

SPDX-License-Identifier: MIT