Speech AI
Turns speech into text and text into natural-sounding speech, and runs voice agents that talk with people live.
Search AirStackWhy teams use it
- Transcribe calls, videos and podcasts automatically
- Add voice to apps, games or accessibility features
- Answer phone calls with a voice agent
- Subtitles and captions in many languages
Who uses it: AI engineer, Software engineer, Video producer, Accessibility specialist, Contact centre manager, Podcast producer. Watch for: Accuracy varies by accent, noise and language; test with your own recordings, and check consent rules for cloned voices. AirStack’s guide to the category.
11 products
| Product | What it says it does | Starting at |
|---|---|---|
| LiveKit Agents | An open source framework and developer platform for building, testing, deploying, scaling, and observing agents in production. | Not observed |
| faster-whisper | Faster Whisper transcription with CTranslate2 | Not observed |
| Pipecat | for building voice and multimodal AI agents | Not observed |
| WhisperX | Automatic Speech Recognition with Word-level Timestamps (& Diarization) | Not observed |
| Cartesia SDK | Not observed | Not observed |
| whisper.cpp | Port of OpenAI's Whisper model in C/C++ | Not observed |
| AssemblyAI SDK | Use the Python or TypeScript SDK to integrate transcription, streaming, and voice agents directly into your own codebase. | Not observed |
| Deepgram SDK | Speech-to-text, text-to-speech, voice agents, and audio intelligence. | Not observed |
| Kokoro | An inference library for Kokoro-82M | Not observed |
| Whisper | Whisper is a general-purpose speech recognition model. | Not observed |
| ElevenLabs SDK | so developers can embed voice into their own applications and workflows | Not observed |
Prices are list prices as written on each pricing page; hover one for its quote. “Not observed” means AirStack has not found it on that product’s pages.
Open models for the job
- argmaxinc/whisperkit-coreml 34,624,804 downloads in 30 days · mit
- jonatasgrosman/wav2vec2-large-xlsr-53-japanese 16,131,153 downloads in 30 days · apache-2.0
- pyannote/speaker-diarization-3.1 6,945,594 downloads in 30 days · mit
- hexgrad/Kokoro-82M 11,706,109 downloads in 30 days · apache-2.0
- audio-cpp/audio.cpp-gguf 8,252,030 downloads in 30 days · other
- coqui/XTTS-v2 5,850,941 downloads in 30 days · other