Voice & AI Framework
ESP-SR
Local speech recognition and edge AI audio processing framework, enabling offline voice commands without internet connection.
- Language
- C / C++
- ESP Speech Recognition Framework
- Category
- Voice & AI Framework
- Main purpose
- Offline voice control with wake word detection, localized command speech-to-intent and acoustic front-end (AFE)
Key features
- WakeNet: high-accuracy wake word engine (e.g., 'Hi ESP', 'Alexa')
- MultiNet: offline speech command recognition for multiple languages
- Acoustic Front-End (AFE): noise suppression, beamforming and echo cancellation (AEC)
- Hardware acceleration on ESP32-S3 via Xtensa vector instructions
- Zero cloud latency and complete user voice privacy
When to use
- Offline voice-controlled home appliances
- Industrial voice terminals in noisy environments
- Privacy-focused smart home gadgets
- Interactive toys and hands-free assistive devices
Code example
Starter snippet in C / C++. Adjust pins and configuration for your board.
ESP-SRC / C++
1