What is API-Route?

What is API-Route

Product position

API-Route is an OpenAI-compatible multi-model AI API. It lets you call supported models through one Base URL and API key instead of maintaining a separate connection for every provider.

The same account brings together current model pricing, account balance, plans, and request logs.

• One Base URL and API key for supported models
• Unified model pricing and availability
• One place for balance and request records

Users, models, and applications

API-Route is designed for AI application development, SaaS integration, and teams that need a unified way to call multiple models.

Models

Supported models can change; use the Pricing page as the current source of truth instead of relying on a fixed list here.

Tools

Current integration guidance covers Codex, Claude Code, LibreChat, VS Code, and other tools listed on the Apps page.

View models and pricing Explore supported applications View the Quickstart

How API-Route works

Four-step workflow

1. Register an account Create the account that owns the balance and API keys. You can register or sign in with a Google, X, or GitHub account.
2. Add balance or buy a plan Use the payment options currently shown on the Top Up page, then choose pay-as-you-go balance or an available plan.
3. Create an API key Create and securely save an enabled key from the API Keys page.
4. Choose a model and call it Use an exact model ID currently available to the key and send a request through the configured endpoint.

API key controls

Each API key can be configured separately, so applications and environments do not have to share the same limits or access rules.

Quota and expiration

Use an unlimited key or set a fixed quota, then choose a specific expiration time or no expiration.

Routing priority

Choose whether the key prioritizes token-priced routes or per-call-priced routes. Available routes still depend on the selected model and current backend status.

Model allowlist

Restrict the key to selected models currently listed by the site. Leaving the list empty adds no model restriction beyond the key group and current listing status.

IP allowlist

Allow one or more IP addresses or CIDR ranges. Leaving the list empty adds no IP restriction.

Key groups and supported models

When key groups are available, create a key for a group, review its pricing, and inspect the models the key can currently use.

Key lifecycle

Edit, enable, disable, or delete each key independently. A new secret is displayed once and should be stored securely.

Manage API keys

One interface

The connection stays simple: the Base URL identifies the endpoint, the API key authorizes the request, and the model ID selects the model. Balance and request logs remain under the same account.

Base URL

The OpenAI-compatible endpoint used by supported clients and SDKs.

API key

The credential created in your account for authenticated API requests.

Model ID

The exact current model identifier selected for a request.

Account and logs

A unified place to review balance, usage, and request records.

Continue to the Quickstart

Multi-route routing and failover

Multi-route design

The same model can have one or more available route nodes. On the Pricing page, “1 route”, “2 routes”, or “3 routes” shows the number of routes currently available for that model.

• Multiple routes can reduce dependence on a single upstream node.
• Not every model has more than one route.
• The current route count is always the value shown on the Pricing page.

Failover and its limits

This multi-provider AI API routing design supports automatic API failover: when the current route errors, times out, or is temporarily unavailable, the system can retry or switch to another available route according to its routing policy.

• An LLM API fallback can reduce the impact of a single-node failure.
• Multiple routes make the AI gateway more fault-tolerant, but do not guarantee zero failures.
• Actual route availability varies by model and time.

User request → Select an available route → Retry or switch on error

Pricing and cost

How API usage is billed

Billing units depend on the selected model. Actual charges are determined by the current Pricing page, account records, and request logs.

Input tokens

Text or context sent to a token-priced model.

Output tokens

Text generated by the model.

Cache fees

Supported models may price cache reads and cache creation separately.

Per-call and media pricing

Images, audio, video, and other media models may charge by call, duration, resolution, or another listed specification.

Balance and plans

Usage can consume account balance or an available plan according to its terms and limits.

Compare model pricing View plans

Why prices can be below official reference pricing

Affordable AI API pricing can come from combining several available upstream suppliers and route nodes. Different suppliers may use different bulk purchasing, partner settlement, and cost structures.

Operating cost

Aggregated request volume, shared infrastructure, and automated routing can lower the cost per call.

Official comparison

Official API pricing is the model provider’s published retail price and is used as a reference for comparison.

Route differences

Routes can differ in price, latency, limits, context capabilities, and availability. Lower AI API pricing does not mean every route has exactly the same characteristics as an official direct connection.

API-Route and other options

Each option has a valid use case. The comparisons below focus on account and integration differences without ranking provider quality.

API-Route and official model APIs

Official APIs remain the direct service from OpenAI, Anthropic, Google, and other providers. API-Route focuses on reducing the operational work of using several providers through one account.

Is API-Route an OpenRouter Alternative?

API-Route and OpenRouter both serve users who want unified access to multiple models. OpenRouter also provides a multi-model interface, OpenAI SDK compatibility, and provider routing, so those capabilities are not unique to API-Route.

The two services have a similar direction. Choose based on payment methods, plan and balance preferences, client configuration, localized support, and the specific models you need.

Get started

Choose a usage mode

AI Chat

Use the site’s web chat with models currently available for that product.

API integration

Connect an application, SDK, script, or supported client through the API.

Independent platform

Use the existing independent-site option when you need a separate platform for your own users.

Next step

Open the detailed setup guide, compare current model costs, or create an account.

View the Quickstart Compare model pricing Create an account