Free account with 25 monthly credits. Pro plan at $7 per month for 1,000 credits. Credit pack one-time purchase of $10 for 1,000 credits. 1 credit per YouTube transcript, 1 credit per audio minute for files and media.
Key features
Streaming speech-to-text for Singlish, English and Mandarin
Natural turn-taking that distinguishes pauses from completion and handles interruptions
Tool calling to pick actions and call custom functions
Grounded responses using instructions and tool results
Local voices in English and Mandarin with natural acknowledgements
Text-to-speech API
Realtime session playground in browser
Python and TypeScript SDKs coming soon
Websocket API for streaming audio
Search transcripts by phrase and timestamp
Export as TXT, SRT, VTT, or JSON
Transcribe YouTube videos with available captions
Upload audio and video files up to 2 GB
Speech-to-text for videos without captions
Transcription API with batch and channel jobs
Keep timestamps with each transcript passage
Transcribe direct links to media files
What makes it different
Built on models trained specifically on local Southeast Asian speech patterns, accents and languages
True real-time architecture with streaming at every layer for immediate responsiveness
Privacy-first design with no call audio stored after sessions end
Handles mixed-language conversations seamlessly within single calls
Searchable transcripts with timestamps tied to video passages
Multiple export formats including SRT and VTT for subtitle editors
Developer API for batch transcription and automation
Speech-to-text transcription for any audio or video file
Tavia Labs
TranscriptLab