Overview
KeyoAPI is an OpenAI-compatible API gateway for developers who want every major model behind one account, one key and one prepaid wallet. Instead of juggling separate invoices and SDKs for GPT-class, Claude-class, Gemini-class and DeepSeek models, you point your existing OpenAI SDK at the KeyoAPI base URL and swap models by changing a single parameter.
Key Features
- OpenAI-compatible: works with the official OpenAI SDKs for Python, Node.js and most third-party clients — change the base URL and you are done.
- Multi-model catalogue: GPT-class, Claude-class, Gemini-class, DeepSeek and open-source models covering text, image, speech, video and OCR workloads.
- Free prototyping: free models are available so you can validate your pipeline before spending anything.
- Transparent pricing: per-model rates published openly, starting at $0.01 per million input tokens.
- One prepaid wallet: top up once, call every model — no subscriptions, no per-vendor billing.
- Cost routing: route batch jobs to cheaper models like gpt-5.6-luna (~$0.10/$0.62 per 1M tokens) and keep premium models for the hard tasks.
Use Cases
- Comparing real OpenAI/ChatGPT API costs across models before committing to one
- Migrating an existing OpenAI project to cheaper or alternative models without rewriting code
- Building AI agents, chatbots, transcription, TTS and image pipelines with a single integration
- Teams that need predictable, prepaid AI spend without enterprise contracts
Getting Started
- Register and top up your wallet.
- Create an API key in the console.
- Set the base URL to https://www.keyoapi.xyz/v1 in your existing OpenAI client.
- Call /v1/chat/completions with any supported model ID.
Pricing & Plans
Pay-as-you-go with no monthly fee. Example rates per 1M tokens (input/output): deepseek-v4.1-flash ≈ $0.01/$0.04, gpt-5.6-luna ≈ $0.10/$0.62, gpt-6-astra ≈ $1.84/$9.19. A typical 50M input + 10M output workload costs about $11.20/month on Luna. The full, regularly updated comparison lives at https://www.keyoapi.xyz/openai-api-pricing — verify live rates on the pricing page before committing.
