top of page

Live websocket STT for just
$0.11 per hour 

gradient green.jpg

$0.11

Cost per hour of live transcription (5x lower than leading STT providers)

gradient blue (1).jpg

6.1%

Word error rate (lower than Deepgram, Assembly, and OpenAI.

gradient green.jpg

270 ms

Median transcription round-trip time, 100-200 ms faster than leading providers.

Built for voice agents

Fennec's websocket streaming has fully tunable voice activity detection (VAD) controls and events, allowing you to lightning fast, interruptable, and turn detection enabled agents.

ai_voice_agent.png

Feature-rich STT

Identify words and phrases that are tough to recognize with STT, and correct them automatically. Works for medical terms, company names, and much more.

Live Contextual Correction

Automatically identify speakers within a transcript. Use our auto lableing feature to identify names as well, no post-processing needed.

Diarization

Automatically format trnascripts based on pause length. No more blocks of text.

Auto Format

Experimental feature for AI agents that only sends a transcript when a complete thought is detected. No more interrupting voice agents!

Thought Detect

Switching has never been easier. 

We provide deep documentation, rich code samples, platform adapters, and an ultra-simple Python SDK. Integrate Fennec in a heartbeat. 

Rapid Integration

With our super-simple Python SDK, REST API, and websocket streaming endpoints, you can start building with Fennec in minutes. 

Built for AI Agents

Fennec ASR's websocket connectiveity has fully configurable VAD for lightning-fast AI Voice Agent Interactions.

Context + Custom Word Injection

Write out contextual instructions and custom words in natural language. Fennec handles recognition and replacement for you. 

Welcome to the new age of live STT.

fennec logo large
bottom of page