Gladia

Gladia is an AI audio infrastructure platform for developers and companies that need to turn live or recorded speech into structured data via a single API.

Key capabilities include:

• Real-time and async speech-to-text with WebSocket support for streaming use cases and sub-300 ms latency for real-time transcription
• Multilingual transcription across 100+ languages, designed for multilingual conversations and global products
• Audio intelligence add-ons such as speaker diarization, sentiment analysis, named entity recognition, word-level timestamps, summarization, and custom vocabulary
• Live translation and enrichment features that help convert calls and meetings into application-ready outputs

Pricing: Freemium, including free access for up to 10 hours per month in the Playground for testing, plus Pay-as-you-Go and Enterprise tiers.

What makes it notable: it combines transcription and enrichment in one API for both streaming and batch workflows, supports telephony-oriented integrations (including SIP use cases), and highlights security and compliance coverage (GDPR, HIPAA, AICPA SOC 2, ISO 27001) along with direct, high-touch support options.