Gladia offers a hosted speech-to-text API that transcribes audio into text with built-in diarization, support for 100 languages and a focus on low latency. The company initially fine-tuned OpenAI’s Whisper model and competes with specialists like AssemblyAI, Deepgram and Speechmatics. Its API is compatible with common telephony stacks and protocols including SIP, VoIP, FreeSwitch and Asterisk. Gladia claims live transcription latency under 300 milliseconds and aims for “batch quality with real-time capabilities.” Over 600 companies use its service, including meeting recorders and note-taking tools such as Attention, Circleback, Method Financial, Recall, Sana and Veed.io. With the new funding it plans to simplify pipelines by integrating audio intelligence and LLM-based tasks in a single API call so customers can get summaries and other outputs without calling third-party LLM APIs.

Total raised

$16M

Funding rounds

1

Latest round

Series A

Latest activity

Oct 2024