Last updated 06-27-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
OnAPI
OnAPI is a unified API gateway that lets developers call GPT, Claude, Gemini, Sora, Veo, and Google's Nano Banana image model through a single key. Swap your SDK base URL to OnAPI's endpoint, plug in an OnAPI key, and keep the rest of your integration unchanged.
The gateway speaks OpenAI-compatible and Anthropic-compatible protocols, so chat completions, Anthropic messages, streaming, function calling, and long-context requests work without a rewrite. OnAPI says new upstream models typically appear within 24 hours of release.
Two routing channels share the same account. The economy channel bills at a 1:1 rate multiplier for everyday volume. The official channel routes through paid keys tied directly to provider accounts at a 1:7.5 multiplier when you need a straighter path to upstream APIs.
OnAPI is aimed at builders shipping AI products, especially where direct provider billing is inconvenient. Top-ups run through Alipay, WeChat, USDT, and USDC. Account balance does not expire, and the site states there is no minimum spend.
The service runs on the open-source New API project (based on One API), which handles model aggregation and cross-format conversion between provider APIs.
One key reaches GPT, Claude, Gemini, Sora 2, Veo, and Nano Banana
OpenAI and Anthropic SDKs work after a base_url swap
Economy channel runs at a 1:1 multiplier, billed pay-as-you-go
Official channel uses direct provider keys at a 1:7.5 multiplier
Streaming, function calling, and long context stay intact
New upstream models added within 24 hours, per OnAPI
One key and endpoint covers text, image, video, and code models from multiple vendors.
Drop-in OpenAI and Anthropic compatibility limits integration rework.
Supports local payment rails including Alipay, WeChat, and crypto top-ups.
Prepaid balance does not expire and carries no stated minimum spend.
Pricing varies by model, so costs are harder to forecast than flat subscriptions.
The public /docs route on onapi.ai returns a page-not-found message; docs live on docs.newapi.pro.
Model marketplace pricing page scrapes returned little detail beyond the homepage copy.
How do I connect OnAPI to my existing code?
OnAPI is OpenAI- and Anthropic-compatible. Point your SDK base URL to https://api.onapi.ai (or https://api.onapi.ai/v1 for OpenAI-style calls) and replace the API key with an OnAPI key from the console. OnAPI's homepage shows sample environment variables for Claude clients.
What models does OnAPI support?
OnAPI lists OpenAI text, image, embedding, Whisper, and TTS models including GPT-5 and the o series; Claude Opus 4, Sonnet 4, and Haiku 4 with 200K context; Gemini 2.5 Pro, Flash, and Flash Thinking with 1M context; Sora 2 for 10s or 15s video; Google Veo for video; and Nano Banana for fast image generation.
What is the difference between OnAPI economy and official channels?
OnAPI's economy channel uses a 1:1 rate multiplier and is positioned for daily volume, agents, and batch jobs. The official channel uses paid keys tied directly to provider billing at a 1:7.5 multiplier. OnAPI states the official route does not rely on chat2api-style relay projects.
Does OnAPI balance expire?
No. OnAPI states that prepaid balance does not expire and that there is no minimum consumption requirement on the account.
How can I pay for OnAPI?
OnAPI supports Alipay, WeChat, USDT, and USDC top-ups from the console. The site also notes a 1:1 recharge display rate where 1 RMB credited shows as 1 USD in the account.
Does OnAPI offer a free tier?
OnAPI's public pricing copy describes pay-as-you-go billing with prepaid balance rather than a standalone free plan. You create an account, fund the balance, and pay per model usage.