LoginSign Up
← Back to all sparks
T

Telnyx

COMMS
Velocity7.5

Programmable voice, messaging, and connectivity APIs for developers.

telnyx.com ↗

Telnyx is bolting every communication channel and every model onto one bill.

email-apivoice-aimodel-choicechannel-consolidationself-hosted-stt
◆Current state
Telnyx runs voice, SMS, and WhatsApp on its own network and has spent this window adding two things: email as a fourth channel, and more model choice inside Voice AI. The Email API entered beta, while Kimi K3 joined Telnyx Inference and GPT-5.4-mini, NVIDIA Parakeet STT, and Fish Audio TTS joined the voice agent stack. Voicemail greetings became configurable per number.
◆Where it's heading
Two threads run in parallel. The channel thread is consolidation — one API and one bill across voice, messaging, and now email — which competes on procurement rather than features. The model thread is optionality: Telnyx is positioning as the place where you pick the STT, LLM, and TTS for a voice agent rather than accepting a vendor's fixed pipeline, with self-hosted options like Parakeet keeping traffic on its own network.
◆Prediction
Email should reach general availability with deliverability tooling attached, since a beta send API without reputation management is not sellable against Postmark or Resend. Expect model additions to continue at roughly this cadence; they are routine here, not directional.

◆Recent moves

  1. 2d ago

    Email API Now Available in Beta

    ⚡ SPARK

    Email becomes a fourth first-party channel alongside voice, SMS, and WhatsApp. It is the only release in this window that changes what Telnyx is rather than what it supports, and it makes the one-API-one-bill pitch a full-stack claim.

    View source ↗
  2. 4d ago

    Kimi K3 Now Available on Telnyx Inference

    A 2.8T-parameter open-source model with a 1M context window and native vision joins Telnyx Inference at $2.70 per million input tokens. Consistent with the model-optionality thread — the value is the menu, not any single addition.

    View source ↗
  3. 9d ago

    GPT-5.4-mini Now Available for Telnyx Voice AI Assistants

    A compact, cheaper LLM option for Voice AI assistants aimed at high-volume workloads. Routine expansion of the model menu, and the cost angle matters more than the capability one for always-on voice agents.

    View source ↗
  4. 12d ago

    Custom Voicemail Greetings Now Available for Telnyx Phone Numbers

    Per-number voicemail greetings via TTS or uploaded audio, configurable in the portal or by API. Ordinary telephony depth, but it closes a gap that shows up immediately when numbers are provisioned programmatically at scale.

    View source ↗
  5. 15d ago

    NVIDIA Parakeet STT Now Available for Telnyx Voice AI

    A self-hosted multilingual STT engine covering 25 European languages with automatic language detection and no third-party API hop. The self-hosted framing is the interesting part: it keeps latency and data inside Telnyx's own network.

    View source ↗
  6. 19d ago

    Fish Audio TTS Now Available for Telnyx Voice AI

    A TTS engine with inline emotion control across 80+ languages joins the voice stack. Another entry in the pick-your-own-pipeline strategy rather than a change of direction.

    View source ↗