--- title: Agent Platform — Telnyx canonical: https://telnyx.com/agent-platform --- # Agent Platform — Telnyx # Agent platform built on owned infrastructure Build, deploy, and scale voice AI agents on the only platform that owns the entire stack, from GPUs that host frontier speech and AI models, to phone numbers, and telephony. One API, one bill, complete ownership. [Build your first agent](https://telnyx.com/sign-up)[View GitHub repo](https://github.com/team-telnyx) WHY TELNYX ## Everything an enterprise voice agent needs, in one place When your agent answers a call, five things have to work in lockstep: the number, the telephony, the speech-to-text, the inference, and the voice that speaks back. Telnyx delivers all five natively, on infrastructure we own — from the GPUs that host your models to the network that carries your call. * ### Calls that get answered Telnyx is a licensed carrier issuing real numbers, with identity and verification handled at the network level. * ### Models on infrastructure we own Frontier speech and language models run on Telnyx GPUs co-located with the call path, not rented from a third party and bolted on. * ### Native orchestration STT, Inference, Voice, Telephony in one pipeline. No webhooks ping-ponging between vendors mid-sentence. OUR INFRASTRUCTURE ## Five vendors is a Frankenstack. One stack is a platform. Stitching Twilio, a TTS vendor, an STT vendor, and an LLM host into one agent builds a Frankenstack: every vendor boundary adds a hop a caller can hear, and every dashboard hides part of the story. Telnyx closes the loop in 200ms on one private network — and AI Insights shows you every conversation, end to end. ### Telnyx Stack 2 hops | 1 vendor PSTNTelnyx Edge PoP → Private IP Network → GPU Cluster (STT + TTS + LLM)PSTN With TelnyxFrankenstack ## 230+ ## COUNTRIES ## $0.05 ## PRICE PER MIN ## 200 ## MILLISECONDS ## A—Level ## CALLER IDENTITY AGENT RUNTIME ## Configure the environment your agents run in Choose the models, voice, and infrastructure your agents will operate on. Once live, agents control the system directly, speaking, routing, and acting without human intervention. Loading... **DEVELOPER QUICKSTART** ## Start from the repo Clone the example agent library, drop in your API key, and you have a working voice agent, config, tool schemas, and README included. `git clone https://github.com/team-telnyx/telnyx-ai-agent-lib-example` Prefer a UI? The portal's AI Assistant Builder gets you to the same place without code. Either path starts with $20 in free credits. HOW TO ## Build your first agent Going from idea to a live agent on a real phone number takes four steps — no code required, unless you want it. Define the agent in the portal's no-code builder or scaffold it from the GitHub repo, pick a voice, attach a number, test it with simulated calls, and go live. Your first agent can take its first call today, with $20 in free credits. [Contact sales](https://telnyx.com/contact-us)[Sign up](https://telnyx.com/sign-up) ### Create Define your agent with a single prompt in the portal's no-code builder, or scaffold it from the GitHub repo with your coding agent. ### Configure Pick a voice, attach a phone number, and connect the tools and knowledge your agent needs. ### Test Run simulated calls in the in-browser tester and refine the agent's behavior before going live. ### Deploy and observe Publish to a real number and watch every conversation end to end with AI Insights. Step 1/4 MIGRATE ## Already running agents somewhere else? From Vapi, Retell, or ElevenLabs — prompts, voices, and tool configs come with you. [Import your agents](https://telnyx.com/sign-up) ElevenLabs ``` from telnyx import Telnyx client = Telnyx(api_key="YOUR_TELNYX_API_KEY") agent = client.ai.assistants.create( name="support-agent", model="Qwen/Qwen3-235B", instructions="You are a helpful support agent.", voice="Telnyx.Ultra.01eaafa9-308a-4276-a017-6ab0cf061b1f", ) ``` Vapi ``` from telnyx import Telnyx client = Telnyx(api_key="YOUR_TELNYX_API_KEY") agent = client.ai.assistants.create( name="support-agent", model="Qwen/Qwen3-235B", instructions="You are a helpful support agent.", voice="Telnyx.Ultra.01eaafa9-308a-4276-a017-6ab0cf061b1f", ) ``` LiveKit ``` from telnyx import Telnyx client = Telnyx(api_key="YOUR_TELNYX_API_KEY") agent = client.ai.assistants.create( name="support-agent", model="Qwen/Qwen3-235B", instructions="You are a helpful support agent.", voice="Telnyx.Ultra.01eaafa9-308a-4276-a017-6ab0cf061b1f", ) ``` Retell AI ``` from telnyx import Telnyx client = Telnyx(api_key="YOUR_TELNYX_API_KEY") agent = client.ai.assistants.create( name="support-agent", model="Qwen/Qwen3-235B", instructions="You are a helpful support agent.", voice="Telnyx.Ultra.01eaafa9-308a-4276-a017-6ab0cf061b1f", ) ``` Bland AI ``` from telnyx import Telnyx client = Telnyx(api_key="YOUR_TELNYX_API_KEY") agent = client.ai.assistants.create( name="support-agent", model="Qwen/Qwen3-235B", instructions="You are a helpful support agent.", voice="Telnyx.Ultra.01eaafa9-308a-4276-a017-6ab0cf061b1f", ) ``` ElevenLabsVapiLiveKitRetell AIBland AI ## Sign up and start building. From first call to global deployment, Telnyx gives you the stack, scale, and support to launch real-time conversational AI with confidence. [Build your first agent](https://telnyx.com/sign-up)[View GitHub repo](https://github.com/team-telnyx) ## Sign up and start building. From first call to global deployment, Telnyx gives you the stack, scale, and support to launch real-time conversational AI with confidence. [Build your first agent](https://telnyx.com/sign-up)[View GitHub repo](https://github.com/team-telnyx) ## FAQ ### What is the Telnyx Agent Platform? The platform for building and running voice AI agents on Telnyx-owned infrastructure — speech recognition, inference, voice, telephony, and messaging in one native stack. ### How fast is it? ### Do I need to write code? ### What does it cost? ### Can I migrate from another platform? [Start building](https://telnyx.com/sign-up)