ROUTER · ONLINEv2026.04 · Seedance 2.0 live

01many. supported models, fragmented

02decide. cost · speed · uptime

one. openai-compatible response03

Multi-model routing infrastructure · AllToken75% off DeepSeek V4 Pro

Many models. One decision.

OpenAI-compatible API for routing text, vision, and video models. Configure model and provider priorities once, then keep billing and cost visibility in one place.

Get API Key /docs

models

providers

uptime target

one

API

Start in minutes

Three steps to route your first request.

Keep your OpenAI SDK pattern, replace the base URL, and start routing across text, vision, and video models through AllToken.

01Create an API key

02Replace your base URL

03Route across models

Create an API key

Generate one AllToken key for your app, environment, or team workflow.

Replace your base URL

Point your existing OpenAI-compatible client to the AllToken endpoint.

Route across models

Choose a model or set routing priorities for cost visibility, provider availability, and capability.

Minimal setupOpenAI SDK

const client = new OpenAI({
  apiKey: process.env.ALLTOKEN_API_KEY,
  baseURL: "https://api.alltoken.ai/v1"
});

Platform features

A practical control layer for multi-model AI.

AllToken focuses on the operating work around model access: routing policy, provider setup, billing, and visibility.

Multi-model routing

Route requests across enabled models with configured priorities and model IDs.

OpenAI-compatible API

Keep familiar SDK patterns and switch providers by changing endpoint and model fields.

Fallback paths

Define provider priorities and fallback paths for provider issues without promising zero downtime.

Unified billing

Use one account balance and one billing layer across supported providers.

Cost and latency visibility

Compare effective tradeoffs before choosing a route policy for production traffic.

Text, vision, image, and video

Reach more than text models through the same API surface as capabilities expand.

Capabilities

One API surface for text, vision, image, and video.

AllToken gives developers a single integration point for model categories that usually require separate provider accounts and setup paths.

Text

Chat and reasoning models

Use leading text models for chat, coding, analysis, and agent workflows.

Vision

Image-aware workflows

Route to models that can inspect images and support multimodal product experiences.

Image

Image generation

GPT Image 2 (and more soon) through the same API surface.

Video

Video generation access

Add video generation workflows without building a separate provider integration.

Seedance 2.0

Seedance 2.0 is a featured video capability.

Seedance 2.0 is highlighted as a video model available through AllToken, with conservative messaging around region availability and competitive pricing.

9:16Product ad clip

16:9Cinematic character shot

1:1Social short concept

Value comparison

Reduce integration overhead before you optimize spend.

AllToken focuses on operational value before spend optimization: fewer provider accounts, clearer billing, and better cost visibility in one place.

One OpenAI-compatible endpointAPI setup

Unified account balanceBilling

Text, vision, image, and videoCapabilities

Fallback paths can be configuredProvider issues

FactorAllTokenDirect providersCredit tools

API setupOne OpenAI-compatible endpointSeparate SDKs, keys, and policiesUsually UI-first and tool-specific

BillingUnified account balanceMultiple invoices and top-upsCredits split by product

Routing controlConfigured model and provider prioritiesImplemented by your teamLimited or unavailable

Provider issuesFallback paths can be configuredYou own retries and alternativesDepends on the tool

FAQ

Questions before you route production traffic.

APIIs it OpenAI-compatible?Swap the base URL and key for most clients.

Yes. AllToken is designed around familiar OpenAI-compatible request patterns so most clients only need a base URL and API key change.

ModelsWhich models are supported?Browse text, vision, and video models in the catalog.

AllToken lists supported text, vision, and video models in the model catalog, including provider and pricing details where available.

RoutingHow does routing work?Choose a model directly or set route priorities.

You can choose model IDs directly or configure priorities that help decide which enabled route should handle a request.

BillingHow is pricing calculated?Pricing follows model, route path, and actual usage.

Pricing depends on the model, provider path, and usage. AllToken helps compare effective costs without relying on unverified lowest-price claims.

VideoCan I use Seedance, vision, and video models?Multimodal models appear where AllToken access is available.

Seedance 2.0 and other multimodal capabilities are positioned as part of the AllToken API surface where available.

FallbackWhat happens if a provider has issues?Fallback behavior follows configured provider priorities.

AllToken can expose configured fallback paths and provider priorities; exact behavior should match the backend routing policy.

Build on AllToken

Start routing with one API key.

Create a key, point your client to AllToken, and compare model paths without wiring every provider by hand.

Get API Key View Docs Explore Models