Release notes

OpenAI GPT-Live, a full-duplex speech-to-speech model, is now a model option in Telnyx Voice AI Assistants, with tool calls handled through delegation.

Telnyx Ultra now supports Urdu and Odia, bringing two widely spoken Indic languages to its voice catalog with no additional configuration.

Self-hosted medical speech-to-text tuned for clinical dialogue, now available across Telnyx Voice API, WebSocket, TeXML, AI Assistants, and the REST API.

Two new smaller, faster LLMs are now available on the Telnyx Inference API, DeepSeek-V4-Flash for cost-efficient workloads and Qwen3.8-27B for low-latency tasks, both served from Telnyx-owned GPU...

Fish Audio TTS is now available for Telnyx Voice AI, adding expressive multilingual speech with inline emotion control to the voice provider catalog.

Telnyx Voice AI now includes Arabic speech-to-text and text-to-speech model options for Arabic-speaking customer interactions.

Inworld Realtime TTS 2 is now available through Telnyx APIs and Voice AI Assistants, extending expressive realtime voices to both standalone API usage and managed agents.

Rime Coda brings higher-quality multilingual text-to-speech to Telnyx, with expanded voice coverage for Hindi, Arabic, Portuguese, Spanish, and more.

Telnyx AI Assistants can now invite participants into a live call and manage turn-taking across multiple speakers. The assistant knows who is speaking, stays silent when people talk to each other,...

Deepgram Flux STT now supports 10 languages on Telnyx Voice AI Assistants, with Multi and Auto language modes for global deployments. Available the same day as Deepgram's GA launch.

Test and compare STT providers directly in Mission Control with the new Speech-to-Text Playground. Transcribe audio from your microphone or upload a file, with no API integration required.

Use dynamic variables to set the voice per call, customizing the assistant voice for each end user without creating separate assistants.