Disclosure in every agent
Agents require a disclosure line and answer truthfully when asked whether they are an AI.
One API for lifelike speech, accurate transcription, natural turn-taking and phone-ready voice agents, running on our own infrastructure.
Pick a sample line or type your own, choose a voice, and press Play.
Performance
Example figures until we publish measured benchmarks.
Developers
Plain HTTPS and JSON. Send text, get audio. Create an agent, open a session, stream audio both ways.
curl -X POST https://api.voxvane.com/v1/text-to-speech/aria \
-H "Authorization: Bearer $VOXVANE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"text": "Hello from VoxVane.", "output_format": "mp3"}' \
--output hello.mp3import os
from voxvane import VoxVane
client = VoxVane(api_key=os.environ["VOXVANE_API_KEY"])
audio = client.text_to_speech("aria", "Hello from VoxVane.", output_format="mp3")
open("hello.mp3", "wb").write(audio)import { writeFile } from "node:fs/promises";
import { VoxVane } from "voxvane";
const client = new VoxVane({ apiKey: process.env.VOXVANE_API_KEY });
const audio = await client.textToSpeech("aria", "Hello from VoxVane.", { outputFormat: "mp3" });
await writeFile("hello.mp3", audio);# 1. Create an agent
curl https://api.voxvane.com/v1/agents \
-H "Authorization: Bearer $VOXVANE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"name": "Front desk",
"voice": "aria",
"instructions": "Answer questions about opening hours and take a message.",
"disclosure": "You are speaking with an AI assistant.",
"turn": {"profile": "balanced"}
}'
# 2. Start a live session (returns a WebSocket URL)
curl -X POST https://api.voxvane.com/v1/agents/agt_123/sessions \
-H "Authorization: Bearer $VOXVANE_API_KEY"import os
import requests
API = "https://api.voxvane.com/v1"
auth = {"Authorization": f"Bearer {os.environ['VOXVANE_API_KEY']}"}
agent = requests.post(f"{API}/agents", headers=auth, timeout=30, json={
"name": "Front desk",
"voice": "aria",
"instructions": "Answer questions about opening hours and take a message.",
"disclosure": "You are speaking with an AI assistant.",
"turn": {"profile": "balanced"},
}).json()
session = requests.post(f"{API}/agents/{agent['id']}/sessions", headers=auth, timeout=30).json()
print("Stream audio to", session["websocket_url"])const API = "https://api.voxvane.com/v1";
const headers = {
Authorization: `Bearer ${process.env.VOXVANE_API_KEY}`,
"Content-Type": "application/json",
};
const agent = await (await fetch(`${API}/agents`, {
method: "POST",
headers,
body: JSON.stringify({
name: "Front desk",
voice: "aria",
instructions: "Answer questions about opening hours and take a message.",
disclosure: "You are speaking with an AI assistant.",
turn: { profile: "balanced" },
}),
})).json();
const session = await (await fetch(`${API}/agents/${agent.id}/sessions`, { method: "POST", headers })).json();
console.log("Stream audio to", session.websocket_url);Products
Use one piece or the whole stack. Every part speaks the same API. Products marked Preview are in private preview; Coming soon means not available yet.
Core voice APIs
Text to speech
Natural, low-latency speech in eight voices, streamed as it renders.
PreviewLearn moreSpeech to text
Accurate transcripts for phone and app audio, live or from a file.
PreviewLearn moreTurn detection
Knows when someone has finished speaking, so replies start fast without interrupting.
PreviewLearn moreConversation model
The reasoning model behind agents, built to check every number against your facts before it says it.
PreviewLearn moreAgents and phone
Real-time agents
Listen, think and speak in one session, with tools, handoff and AI disclosure built in.
PreviewLearn morePhone and telephony
Phone numbers, SIP, transfers and voicemail for agents that answer real calls.
PreviewLearn moreAgent template
A ready-made phone receptionist: answers, takes complete messages and transfers on request.
PreviewLearn moreAgent testing
Simulated callers and scored test runs, so you know an agent works before it takes a real call.
PreviewLearn morePlatform
One API, many providers
One API across speech and model providers, with automatic fallback and a cost record for every call.
PreviewLearn moreTranscripts and summaries
Clean transcripts, short summaries and redaction of sensitive details after every conversation.
PreviewLearn morePrivacy tier
A HIPAA-ready privacy tier. BAA available on request (coming soon).
Coming soonLearn moreConsented voice cloning
Custom voices from a speaker who has given recorded consent, watermarked on every use.
Coming soonLearn moreOn-premises
The VoxVane stack on hardware in your own building. Contact sales.
Coming soonLearn morePhone agent demo
An example use case: an after-hours receptionist for a fictional plumbing company. Press play and watch the transcript and structured output fill in live.
Voices
Press play to hear each voice introduce itself.
US English
US English
US English
US English
US English
US English
UK English
UK English
Use cases
Voice Agents · template
An agent that answers a business line, takes a complete message and hands off to a person on request.
ExploreVoice Agents
Order status, account questions and triage, with a clean transcript for your team.
ExploreVoice + Listen
Give your product a voice: stream speech in, stream speech out, over one WebSocket.
ExploreVoice
Turn articles, lessons and notifications into natural audio on demand.
ExploreResponsible by default
Agents require a disclosure line and answer truthfully when asked whether they are an AI.
A custom voice needs the speaker's recorded consent and is watermarked on every use.
We build on open models with commercial licences, run them ourselves, and credit every project.
The API is in private preview. Read the quickstart, then ask for a key.