{"company":{"name":"Soniox","slug":"soniox","website":"https://soniox.com","category":"ai-ml"},"question":"What has Soniox shipped recently?","answer":"In the last 30 days, Soniox shipped 12 tracked updates. The most recent was \"Real-time speech translation API\" on 2026-11-08.","window":{"days":30,"updateCount":12,"returned":12},"generatedAt":"2026-11-22T08:03:46.137Z","updates":[{"title":"Real-time speech translation API","summary":"Soniox introduced a real-time speech translation API supporting 60+ languages and 3,600+ language pairs, enabling one-way or two-way translation during live speech. Pricing starts at $0.18/hour for real-time use, with async options at $0.16/hour. The API preserves names, numbers, and domain-specific terms while streaming translations before sentences complete.","date":"2026-11-08","dateIsEstimated":true,"signalType":null,"signalTypeLabel":null,"sourceUrl":"https://soniox.com/speech-translation","publisher":"soniox.com"},{"title":"Data residency","summary":"Soniox introduced data residency controls allowing users to select where their audio, transcripts, and metadata are processed and stored—US, EU, Japan, or India—via regional API domains. System data may still be processed outside the chosen region, but user content remains within it.","date":"2026-11-08","dateIsEstimated":true,"signalType":"feature_update","signalTypeLabel":"Feature","sourceUrl":"https://docs.soniox.com/data-residency","publisher":"docs.soniox.com"},{"title":"Text-to-speech API for voice agents and voice bots","summary":"Soniox updated its Text-to-Speech API to better serve AI voice agents and bots, adding streaming input/output for real-time responses, accurate readback of structured data, and support for 60+ languages with consistent voice quality. The API is designed to handle interruptions and integrate seamlessly with Soniox’s Speech-to-Text API for full conversational loops.","date":"2026-11-08","dateIsEstimated":true,"signalType":null,"signalTypeLabel":null,"sourceUrl":"https://soniox.com/text-to-speech/use-cases/voice-agents","publisher":"soniox.com"},{"title":"Text-to-speech API for video voiceover and localization","summary":"Soniox enhanced its Text-to-Speech API to support expressive video voiceovers and multilingual localization with 200+ studio-quality voices, voice cloning, and 60+ languages. Features include inline audio tags for emotion control, character-level timestamps for alignment, and production-ready audio formats up to 48 kHz.","date":"2026-11-08","dateIsEstimated":true,"signalType":null,"signalTypeLabel":null,"sourceUrl":"https://soniox.com/text-to-speech/use-cases/video-voiceover","publisher":"soniox.com"},{"title":"Text-to-speech API for audiobook and narration products","summary":"Soniox enhanced its Text-to-Speech API with voice cloning from short reference clips, inline audio tags for emotion and pacing, and support for 60+ languages with consistent narrator identity across translations. The API now targets long-form audiobook and narration workflows with programmatic control over audio output and character-level timestamps.","date":"2026-11-08","dateIsEstimated":true,"signalType":null,"signalTypeLabel":null,"sourceUrl":"https://soniox.com/text-to-speech/use-cases/audiobooks","publisher":"soniox.com"},{"title":"Text-to-speech API for e-learning, training, and AI tutors","summary":"Soniox enhanced its Text-to-Speech API to better serve e-learning, training, and AI tutors by adding multilingual support across 60+ languages, voice cloning from short recordings, and real-time streaming for responsive interactions. The API now supports character-level timestamps, audio tags for emotion control, and integration with Soniox Speech-to-Text for two-way voice experiences.","date":"2026-11-08","dateIsEstimated":true,"signalType":null,"signalTypeLabel":null,"sourceUrl":"https://soniox.com/text-to-speech/use-cases/e-learning","publisher":"soniox.com"},{"title":"Text-to-speech API for games, NPCs, and live interactive characters","summary":"Soniox updated its Text-to-Speech API to support direct emotion control via audio tags, voice cloning for consistent character identity, and real-time streaming for dynamic dialogue in games and AI characters. The API now handles scripted and runtime-generated speech with up to 5 concurrent streams per connection and supports 60+ languages with in-region processing.","date":"2026-11-08","dateIsEstimated":true,"signalType":null,"signalTypeLabel":null,"sourceUrl":"https://soniox.com/text-to-speech/use-cases/games","publisher":"soniox.com"},{"title":"Voice dictation API for fast, accurate, and multilingual voice typing","summary":"Soniox updated its Speech-to-Text API to better support voice dictation with low-latency streaming, provisional and finalized transcription, and native-speaker accuracy across 60+ languages. The API now handles structured details like names, numbers, and dates while offering in-region processing and privacy compliance.","date":"2026-11-08","dateIsEstimated":true,"signalType":"feature_update","signalTypeLabel":"Feature","sourceUrl":"https://soniox.com/speech-to-text/use-cases/voice-dictation","publisher":"soniox.com"},{"title":"Supported languages","summary":"Soniox announced its Text-to-Speech AI now supports over 60 languages using a single unified AI model, enabling natural and expressive speech generation. All languages are available in standard and neural variants, with voice cloning and language mixing features supported across the expanded set.","date":"2026-11-08","dateIsEstimated":true,"signalType":"feature_update","signalTypeLabel":"Feature","sourceUrl":"https://docs.soniox.com/tts/concepts/supported-languages","publisher":"docs.soniox.com"},{"title":"Speaker diarization API for real-time and recorded audio","summary":"Soniox launched speaker diarization in its speech-to-text API, labeling up to 15 speakers in real-time or recorded audio across 60+ languages with no additional cost. The feature streams speaker labels per token and supports one-parameter activation, competing on price and privacy with providers like OpenAI and Google.","date":"2026-11-08","dateIsEstimated":true,"signalType":"feature_update","signalTypeLabel":"Feature","sourceUrl":"https://soniox.com/speech-to-text/features/diarization","publisher":"soniox.com"},{"title":"Real-time generation","summary":"Soniox introduced real-time Text-to-Speech generation through a WebSocket API that streams audio back as text arrives, enabling sub-second latency for LLM-driven workloads. The two-way stream allows concurrent audio generation while text is still being processed, eliminating buffering delays for voice agents and interactive assistants.","date":"2026-11-08","dateIsEstimated":true,"signalType":null,"signalTypeLabel":null,"sourceUrl":"https://docs.soniox.com/tts/rt/real-time-generation","publisher":"docs.soniox.com"},{"title":"Real-time transcription","summary":"Soniox introduced real-time transcription via WebSocket API, supporting over 60 languages with low latency and high accuracy. The system streams provisional and final tokens, allowing instant feedback and stable output, and supports auto-detected audio formats without configuration. Final tokens are confirmed once and never change.","date":"2026-11-08","dateIsEstimated":true,"signalType":"feature_update","signalTypeLabel":"Feature","sourceUrl":"https://docs.soniox.com/stt/rt/real-time-transcription","publisher":"docs.soniox.com"}],"attribution":{"source":"Spyingbee","url":"https://spyingbee.com/updates/soniox","citation":"Spyingbee, \"Soniox updates\", https://spyingbee.com/updates/soniox (retrieved 2026-11-22)"}}