← Back to all sparks
T

Telnyx

COMMS
Velocity7.5

Programmable voice, messaging, and connectivity APIs for developers.

Telnyx is stitching every new STT, TTS, and LLM into one on-network voice AI stack.

voice-aimodel-integrationsspeech-to-textagent-workflowslow-latencyon-network-inference
Current state
Telnyx has turned its AI Assistant layer into a model marketplace, onboarding speech-to-text, text-to-speech, and LLM options at a steady clip. The differentiator is on-network inference: models run on Telnyx infrastructure instead of being stitched across third-party vendors. Recent weeks added Kimi K2.6, GPT-5.4, several STT engines, new TTS voices, and conversation-flow tooling.
Where it's heading
The product is moving from a voice API with bolt-on AI toward a full agent-building platform where customers pick models per step and route conversations through workflow logic. Each release either widens model choice or tightens latency, via anchorsites and on-network processing. The consistent pattern is breadth of integrations plus control over the orchestration layer.
Prediction
Expect continued rapid onboarding of new frontier LLMs and STT/TTS engines, alongside deeper workflow and routing features that make the assistant builder more programmable.

Recent moves

  1. 1d ago

    Minimax M3 is now available on Telnyx Inference

    View source ↗
  2. 2d ago

    Kimi K2.6 Now Available for Telnyx AI Assistants

    ⚡ SPARK

    Adding Moonshot's Kimi K2.6 continues Telnyx's pattern of onboarding frontier LLMs into its assistant layer the moment they are viable, this time targeting agentic-coding workloads.

    View source ↗
  3. 10d ago

    Conversation Workflows Now Available for Telnyx AI Assistants

    ⚡ SPARK

    Workflow nodes with conditional routing and per-step model and voice overrides mark Telnyx's shift from single-prompt assistants to programmable multi-step agents.

    View source ↗
  4. 10d ago

    Soniox STT Now Available for Telnyx Voice AI

    Soniox STT joins the roster, extending the multilingual, streaming transcription options available inside Telnyx voice workflows.

    View source ↗
  5. 10d ago

    Deepgram Flux Multilingual STT Now Available for LiveKit Agents on Telnyx

    Deepgram Flux for LiveKit agents adds automatic language detection and low-latency turn-taking, reinforcing Telnyx's strategy of offering multiple best-fit STT engines rather than one default.

    View source ↗
  6. 18d ago

    Rime Coda TTS now available on Telnyx

    Rime Coda adds 236 voices across 8 languages with on-network processing, deepening the TTS side of the same multi-vendor, low-latency playbook.

    View source ↗