Route Key: OpenAI-Compatible AI API Gateway

Route Key is a hosted OpenAI-compatible AI API gateway for multi-model routing, AI API pricing comparison, provider failover, and usage visibility.

Use one API integration to connect supported GPT, Claude, Gemini, DeepSeek, Grok, Qwen, and other model providers. Keep familiar OpenAI-compatible request formats while selecting routes by model capability, price, health, latency, and endpoint support.

• Website: https://routekey.ai/
• Developer docs: https://routekey.ai/en/pc/docs
• Models and pricing catalog: https://routekey.ai/en/pc/models
• Integration guide: https://routekey.ai/en/pc/docs/integrations
• Practical guides: https://routekey.ai/en/blog
• Organization: https://github.com/RouteKey

What Route Key Provides

• OpenAI-compatible API access through one base URL
• Multi-model routing and explicit model selection
• Cost-aware, capability-aware, and health-aware route policies
• Provider and channel failover signals
• API key management and Usage Logs
• Token, latency, retry, status, and cost visibility
• Coding-agent integrations for Codex, Claude Code, Gemini CLI, Cursor, and compatible tools

Model availability, endpoint support, pricing signals, and account-group access can change. Use the live Models catalog for the current model ID and pricing data before production traffic.

Quick Start

1. Create an API key

Create a Route Key account and API key in the Dashboard. Keep a separate key for each environment or coding agent so Usage Logs and rotation remain clear.

2. Set the compatible base URL

https://api.routekey.ai/v1

Use the exact model ID shown in the Models catalog. Do not put an API key in a repository, issue, screenshot, shell history shared with a team, or client-side application bundle.

3. Send a first request

export ROUTEKEY_API_KEY="your_routekey_api_key"
export ROUTEKEY_MODEL="MODEL_ID_FROM_CATALOG"

curl https://api.routekey.ai/v1/chat/completions \
  -H "Authorization: Bearer ${ROUTEKEY_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "{\"model\":\"${ROUTEKEY_MODEL}\",\"messages\":[{\"role\":\"user\",\"content\":\"Reply with one short sentence.\"}]}"

Copy the same request from examples/curl/quickstart.sh, or use the SDK examples:

• Python OpenAI SDK
• Node.js OpenAI SDK

Developer Guides

• OpenAI-compatible API quickstart
• Multi-model AI API routing
• AI API pricing comparison method
• Coding agents with an OpenAI-compatible API

OpenAI SDK Compatibility

Most OpenAI-compatible clients need only an API key, the Route Key base URL, and an enabled model ID. Keep the request shape from your existing client while testing the endpoints and features your application actually uses.

Test these separately when applicable:

• Chat completions and streaming
• Tool calls and structured output
• Vision, audio, or other multimodal inputs
• Timeouts, retries, and cancellation
• Context limits and token accounting

The API integration guide contains setup profiles for common applications and coding agents. The developer docs contain request and endpoint references.

Routing and Cost Control

Route Key can keep the application API surface stable while a platform team evaluates different model or provider routes. A safe rollout normally uses one default model, one eligible lower-cost alternative, and one tested fallback.

Compare more than a headline token price:

• Input, output, and cached-input tokens
• Context and endpoint support
• Model capability and quality on your task set
• Latency, rate limits, errors, and retries
• Cost per successful task rather than cost per request

Read the AI API pricing comparison guide and inspect the live pricing catalog before changing production policy.

Coding Agents

OpenAI-compatible coding tools usually need a base URL, a secret API key, and an exact model ID. Use a dedicated key for each local tool, workspace, CI job, or production automation environment.

See the coding-agent guide and the official integration guide for Codex, Claude Code, Gemini CLI, Cursor, and other compatible clients.

Security Notes

• Treat API keys as secrets and rotate exposed keys immediately.
• Use environment variables or the tool's protected secret store.
• Never commit .env files, keys, access tokens, or private prompts.
• Start with a small request and a limited balance or quota.
• Review Usage Logs after a new model, route, or client is enabled.
• For a security report, follow SECURITY.md instead of opening a public issue with sensitive details.

Frequently Asked Questions

Do I need to rewrite an OpenAI SDK integration?

Usually no. Configure the SDK with the Route Key base URL and API key, then test the endpoints and features your application uses. Provider-specific options may still need a compatibility review.

Does Route Key silently replace my requested model?

Use explicit model IDs and review the selected model, endpoint, and route details in Usage Logs. Routing and fallback behavior should be part of the policy you configure and verify, not an assumption hidden from operators.

Where can I compare AI API pricing?

Use the Models catalog for current model, vendor, endpoint, group, and pricing signals. Prices and availability can change, so confirm the current catalog before a large purchase or rollout.

How do I get help?

Start with the integration guide and developer docs. Use the Route Key website contact panel for account, billing, or technical support.

About Route Key

Route Key is operated by ROUTEKEY TECHNOLOGY LIMITED. The canonical product website is https://routekey.ai/.

This repository is a public integration reference and documentation hub. It is not a distribution of Route Key's hosted service or private backend source. No reuse license is granted unless a license file is added to this repository.

Contributing

Documentation and safe example improvements are welcome. Read CONTRIBUTING.md before opening a pull request. Never include credentials or private customer data in an issue, pull request, example, or screenshot.