Inference Engine v2.4.0

[ start_inference ]< documentation / >
Global avg: 2.3ms p50
Scroll
Section 01

The Arsenal

Production-grade endpoints. No warm-up. No queueing. Inference at the speed of thought.

View all models→
IMAGE GENERATION|340ms

Flux Schnell

cURL Request
curl https://api.tuen.fun/v1/images/generations \
  -H "Authorization: Bearer $KEY" \
  -d '{"model":"flux-schnell","prompt":"cyberpunk city"}'
LLM|12ms/token

Llama 3.1 70B

cURL Request
curl https://api.tuen.fun/v1/chat/completions \
  -H "Authorization: Bearer $KEY" \
  -d '{"model":"llama-3.1-70b","messages":[{"role":"user","content":"Hello"}]}'
TEXT TO SPEECH|89ms

VibeVoice

cURL Request
curl https://api.tuen.fun/v1/audio/speech \
  -H "Authorization: Bearer $KEY" \
  -d '{"model":"vibevoice","input":"Hello world","voice":"alloy"}'
TRANSCRIPTION|120ms

Cohere Transcribe

cURL Request
curl https://api.tuen.fun/v1/audio/transcriptions \
  -H "Authorization: Bearer $KEY" \
  -F file=@audio.mp3 \
  -F model="cohere-transcribe"
VIDEO GENERATION|2.4s

Nucleus Video

cURL Request
curl https://api.tuen.fun/v1/video/generations \
  -H "Authorization: Bearer $KEY" \
  -d '{"model":"nucleus","prompt":"robot walking"}'
Section 02

Architecture

Bare-metal GPU clusters. Custom inference kernels. No virtualization overhead.

Throughput (TPS)
p50: 2,847p99: 3,102
Latency (p50)
Target: <5ms✓ Within SLO
Active Clusters5 regions
2,048
Total GPUs
H100 / A100 mix
99.997%
Uptime
Last 30 days
0ms
Cold Boot
Always warm
12.4M
Throughput
Requests/day
Section 03

The Playground

Write code. Watch it execute in real-time. Every keystroke drives the machine.

playground.ts
1import { TuenClient } from "@tuen/sdk";
2 
3const client = new TuenClient({
4 apiKey: process.env.TUEN_API_KEY,
5});
6 
7async function generateImage() {
8 const result = await client.images.generate({
9 model: "flux-schnell",
10 prompt: "A cyberpunk city at night",
11 size: "1024x1024",
12 });
13 
14 return result.url;
15}
16 
17generateImage().then(console.log);
TypeScriptUTF-8
Section 04

Wall of Fame

Builders shipping inference at scale

AL
@alex_dev
Active Shipper
SA
@sarah_ml
Active Shipper
CH
@chen_ai
Active Shipper
JO
@jordan_viz
Active Shipper
MA
@maria_nlp
Active Shipper
KA
@kai_gpu
Active Shipper
TA
@taylor_llm
Active Shipper
RI
@rio_deploy
Active Shipper
NI
@nina_opt
Active Shipper
LE
@leo_infra
Active Shipper
MA
@maya_data
Active Shipper
NO
@noah_edge
Active Shipper
ZO
@zoe_tune
Active Shipper
MA
@max_speed
Active Shipper
AV
@ava_cloud
Active Shipper
AL
@alex_dev
Active Shipper
SA
@sarah_ml
Active Shipper
CH
@chen_ai
Active Shipper
JO
@jordan_viz
Active Shipper
MA
@maria_nlp
Active Shipper
KA
@kai_gpu
Active Shipper
TA
@taylor_llm
Active Shipper
RI
@rio_deploy
Active Shipper
NI
@nina_opt
Active Shipper
LE
@leo_infra
Active Shipper
MA
@maya_data
Active Shipper
NO
@noah_edge
Active Shipper
ZO
@zoe_tune
Active Shipper
MA
@max_speed
Active Shipper
AV
@ava_cloud
Active Shipper
AV
@ava_cloud
Active Shipper
MA
@max_speed
Active Shipper
ZO
@zoe_tune
Active Shipper
NO
@noah_edge
Active Shipper
MA
@maya_data
Active Shipper
LE
@leo_infra
Active Shipper
NI
@nina_opt
Active Shipper
RI
@rio_deploy
Active Shipper
TA
@taylor_llm
Active Shipper
KA
@kai_gpu
Active Shipper
MA
@maria_nlp
Active Shipper
JO
@jordan_viz
Active Shipper
CH
@chen_ai
Active Shipper
SA
@sarah_ml
Active Shipper
AL
@alex_dev
Active Shipper
AV
@ava_cloud
Active Shipper
MA
@max_speed
Active Shipper
ZO
@zoe_tune
Active Shipper
NO
@noah_edge
Active Shipper
MA
@maya_data
Active Shipper
LE
@leo_infra
Active Shipper
NI
@nina_opt
Active Shipper
RI
@rio_deploy
Active Shipper
TA
@taylor_llm
Active Shipper
KA
@kai_gpu
Active Shipper
MA
@maria_nlp
Active Shipper
JO
@jordan_viz
Active Shipper
CH
@chen_ai
Active Shipper
SA
@sarah_ml
Active Shipper
AL
@alex_dev
Active Shipper
Section 05

Deploy Now

One command. Instant access. No credit card required to start.

tuen-cli
>register --email

Press Enter to execute. Sound on recommended.