Gladia

About Gladia

Gladia is an audio transcription and intelligence provider offering a speech-to-text API for both real-time and asynchronous (pre-recorded) transcription of audio and video. Beyond transcription, it provides audio intelligence features including diarization, translation, subtitles, custom vocabulary and spelling, named entity recognition, PII redaction, sentiment analysis, chapterization, and summarization, along with tools to apply LLMs directly to audio content.

The product is aimed at developers building applications that require speech processing, such as voice AI agents, meeting recorders, and telephony integrations. It offers an SDK, a CLI, and a playground for testing, along with documentation covering language detection, code switching, and supported languages.

Gladia integrates with third-party platforms including Pipecat, LiveKit, Vapi, Recall, Meeting BaaS, Attendee, Twilio, VideoSDK, Composio, Zapier, Make, and n8n. It also provides migration guides for users moving from Deepgram or AssemblyAI, and documents concurrency/rate limits, multi-channel audio support, supported file types and durations, and data retention policies.

Key features

• Real-time and asynchronous speech-to-text transcription
• Automatic language detection and code switching across supported languages
• Audio intelligence tools including diarization, translation, and subtitles
• PII redaction, sentiment analysis, and named entity recognition
• Chapterization and summarization of audio content
• Integrations with Pipecat, Livekit, Vapi, Twilio, Zapier, Make, and n8n

Frequently asked questions