SpeechRecognition
github.com
- Free plan
- Windows
- macOS
- Linux
- API
- Self-hosted
What SpeechRecognition is
SpeechRecognition is a free Python library that connects applications to multiple speech-recognition engines and APIs, with online and offline options. Supported choices include CMU Sphinx, Google Speech Recognition, Google Cloud Speech API, Wit.ai, Microsoft Azure Speech, Houndify, IBM Speech to Text, Vosk, Whisper, Groq Whisper, and Cohere Transcribe. CMU Sphinx, Snowboy Hotword Detection, Vosk, and Whisper are identified as offline-capable. Examples cover microphone recognition, audio-file transcription, saving audio data, extended results, ambient-noise calibration, and background listening. Installation can use pip install SpeechRecognition or the PyPI source distribution. Python 3.10 or later is required. PyAudio 0.2.11 or later is needed only for microphone input. Google Cloud use requires local credentials, while OpenAI and Groq recognizers require environment API keys. The OpenAI transcription recognizer supports compatible self-hosted endpoints such as vLLM and Ollama. The library uses the 3-clause BSD license; included FLAC binaries use GPLv2. Its Free/open-source library plan costs nothing.
Who it suits
SpeechRecognition suits Python developers who need to switch among online services, offline engines, or self-hosted compatible endpoints. It also fits projects handling microphone input or recorded audio files.
Good to know, and what to watch
- Supports many recognition engines
- Includes offline-capable options
- Handles microphone and audio files
- Supports self-hosted compatible endpoints
- Free under an open-source license
- Requires Python 3.10 or later
- PyAudio needed for microphone input
- Cloud services require credentials or API keys
- Google Cloud Speech V2 is unsupported
Plans
Every plan and priceFree/open-source library
Free plan
The verdict
SpeechRecognition offers one Python interface across a broad set of recognition engines, with offline and cloud paths. Review dependency, credential, and Google API-version requirements before selecting an engine.
What the maker publishes
- Cloud credentials
- The README requires local Google credentials for Google Cloud Speech use and environment API keys for OpenAI and Groq recognizers. source
- Founded
- 2014 source
- Google API limit
- The README says the library supports Google Cloud Speech API V1 and does not support V2. source
- Input and output
- Examples cover recognizing microphone input, transcribing audio files, saving audio data, extended results, ambient-noise calibration, and background microphone listening. source
- Installation
- The README gives `pip install SpeechRecognition` as the easiest installation method and also describes installing from the PyPI source distribution. source
- License
- The library is available under the 3-clause BSD license; included FLAC binaries are GPLv2-licensed. source
- Microphone requirement
- PyAudio 0.2.11 or later is required only for microphone input; the rest of the library works without it. source
- Offline operation
- The README identifies CMU Sphinx, Snowboy Hotword Detection, Vosk, and Whisper as working offline. source
- Purpose
- SpeechRecognition is a Python library for speech recognition with support for several engines and APIs, online and offline. source
- Python requirement
- Python 3.10 or later is required to use the library. source
- Recognition engines
- Supported options include CMU Sphinx, Google Speech Recognition, Google Cloud Speech API, Wit.ai, Microsoft Azure Speech, Houndify, IBM Speech to Text, Vosk, Whisper, Groq Whisper, and Cohere Transcribe. source
- Self-hosted endpoints
- The OpenAI transcription recognizer supports OpenAI-compatible self-hosted endpoints such as vLLM and Ollama. source
- Support
- The project asks users to report bugs and suggestions through its GitHub issue tracker. source
Checked against the maker's pages on 5 Oct 2026.