Model infrastructure for AI app teams

Every model your app needs. One API to ship it.

Keep one OpenAI-compatible interface while NeoRouter handles provider differences, billing, monitoring, guardrails, and automatic fallback across language, voice, image, and video.

One billAutomatic fallbackEnd-to-end traces

Get API key Read the docs

OpenAI-compatible. Integrate in minutes.

quickstart.js

// Keep your SDK. Change the base URL.
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.neorouter.app/v1",
  apiKey: process.env.NEOROUTER_API_KEY
});

const response = await client.chat.completions.create({
  model: "openai/gpt-5",
  messages: [{ role: "user", content: "Hello" }]
});

SWITCH MODEL

POST/v1/chat/completions 482 ms200 OK

All systems operational API EDGE · HKG

One endpoint, leading models across every modality

TEXTGPT-5ClaudeGeminiQwenMiniMax

VOICEElevenLabsDoubaoMiniMax Speech

IMAGEImagenFLUX

VIDEOVeoSeedance

THE MODEL NETWORK

One API contract. Every model behind it.

Integrate once. NeoRouter normalizes APIs, authentication, streaming, errors, and response formats across OpenAI, Anthropic, Gemini, and more.

01 · INPUT Your appPOST /v1/responses

UNIFIED ENDPOINTNeoRouter APIapi.neorouter.app

02 · LIVE PROVIDERS

TEXTGPT · Claude · Gemini · Qwen24

VOICEElevenLabs · Doubao · MiniMax12

IMAGEImagen · FLUX · Recraft15

VIDEOVeo · Seedance · Kling9

2.4MRequests today

99.99%Gateway uptime

THE OPERATIONS LAYER

Connect the models. Hand us the operational drag.

NeoRouter keeps the model layer flexible and observable, so a small AI app team can operate like a much larger platform team.

CONNECT

Use any model

Every provider brings a different SDK, auth flow, and response format.

Keep one OpenAI-compatible contract across language, voice, image, and video models.

OpenAI · Anthropic · Gemini · more

STAY ONLINE

Keep requests moving

Upstream limits and incidents quickly become errors your users can see.

Retry and fail over across healthy providers using live availability signals.

Retry · fallback · health

CONTROL COST

Control every call

Multiple accounts, prices, and quotas make spend hard to control.

Use one balance and set routing policies by budget, model price, or request limits.

Budgets · limits · one bill

OBSERVE

See every request

Fragmented logs make latency, failures, and spend difficult to explain.

Trace latency, cost, status, and the full fallback chain from one calm view.

Logs · latency · cost

PROTECT

Protect model traffic

Leaked keys, abusive input, and usage spikes can multiply model costs.

Centralize keys, rate limits, content policies, and abnormal traffic controls.

Rate limits · guardrails · keys

CHANGE SAFELY

Switch without shipping

Models and providers change too quickly for hard-coded production logic.

Adjust models, weights, and routing policy in the control plane without redeploying your app.

Policies · weights · zero redeploy

ROUTING ENGINE

Choose the outcome. We choose the route.

Try the live policy controls. NeoRouter evaluates price, quality, latency, and availability before each request leaves the gateway.

ROUTER · LIVEpolicy/default-v2···

01 · APP REQUESTGenerate response

modality: text
region: auto

active policyNEOROUTER

GPT-5Live

412 ms

Best live latency in region

SELECTED ROUTE

Claude SonnetLive

92 eval

Highest evaluation score for this task

Qwen MaxLive

$0.8 / 1M

Lowest projected token cost

Gemini 2.5 ProLive

99.99 %

Healthiest independent fallback route

Routing decisions happen before your user notices.policy: fastest · decision: 8ms

DEVELOPER EXPERIENCE

Change one URL. Keep your workflow.

Use the SDKs and patterns your team already knows. NeoRouter is OpenAI-compatible and works naturally across web, iOS, Android, and backend services.

✓Typed model IDs across modalities
✓Consistent errors and streaming
✓Server and client SDK examples

Read the docs →

import OpenAI from "openai";

const neo = new OpenAI({
  baseURL: "https://api.neorouter.app/v1",
  apiKey: process.env.NEOROUTER_API_KEY
});

const result = await neo.responses.create({
  model: "anthropic/claude-sonnet",
  input: "Design a quiet onboarding flow."
});

import OpenAI

let config = OpenAI.Configuration(
  token: ProcessInfo.processInfo.environment["NEOROUTER_API_KEY"]!,
  host: "api.neorouter.app"
)

let neo = OpenAI(configuration: config)
let result = try await neo.chats(query: .init(
  messages: [.user(.init(content: .string("Hello")))],
  model: "openai/gpt-5"
))

from openai import OpenAI

neo = OpenAI(
  base_url="https://api.neorouter.app/v1",
  api_key=os.environ["NEOROUTER_API_KEY"]
)

result = neo.responses.create(
  model="google/gemini-pro",
  input="Summarize this document."
)

curl https://api.neorouter.app/v1/responses \
  -H "Authorization: Bearer $NEOROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "minimax/text-01",
    "input": "Write a launch note."
  }'

api.neorouter.appOpenAI compatible

OBSERVABILITY INCLUDED

See every model call, clearly.

Monitor latency, usage, cost, and fallback behavior from one calm operational view.

PROJECT / PRODUCTIONOverview

Requests24,892+18.4%

Success rate99.97%+0.12%

P95 latency684 ms−42 ms

Spend today$38.21$0.0015 / req

REQUEST VOLUME

Requests Fallbacks

LIVE REQUESTS

MODELROUTELATENCYCOSTSTATUS

Oopenai/gpt-5OpenAI · US-W482 ms$0.0042Complete

Aanthropic/sonnetAnthropic · US-E614 ms$0.0038Complete

Mminimax/speech-02MiniMax · APAC721 ms$0.0081Fallback

SIMPLE PRICING

Pay for what your app uses.

No platform subscription. Add credits and access every supported provider through one balance.

PAY AS YOU GO01

$0platform fee to start

✓Access to every modality
✓Unified usage and billing
✓Routing and fallback controls
✓No long-term commitment

Start building↗

Provider usage is billed at transparent model rates.

SHIP THE MODEL, NOT THE INTEGRATION

Build with every model through one API.

Your next model is one request away.

Get API key↗Read the docs→