I build a free multimodal AI API that works with any OpenAI SDK

I got tired of juggling three different providers to prototype one multimodal app — a key for chat, another for images, another for video, each with its own SDK and a credit card requirement before you can test anything.

So I built NFlashX: one API key for chat, image generation and video, behind a single OpenAI-compatible interface. It's live, and the free tier doesn't ask for a card.

Free tier and pricing

| Model | Output | Rate limit | Price (per 1M tokens) |
| NFlashX 1.0 Flash | 4K | 2 req/min on free |
| NFlashX 1.0 Pro | 8K | 30 req/min | $0.40 in / $0.70 out |
| NFlashX Vision 1.0 Flash | image | — | $0.01 per image |

New accounts start with $1.00 in credit — enough to run a real test, not just a hello-world. Limits are per key, rates are published, and everything lives under a stable /v1.

Why one more API?

Partly selfish: I wanted a sane way to prototype multimodal features without wiring three SDKs. Partly practical: not everyone can put an international card on file, so top-ups also work through local QR transfer and USDT.

It's a young, small project — the first 100 accounts are open.

Feedback welcome

• Is the request/response format close enough to OpenAI for true drop-in use?
• What would make you pick this over a big provider for a side project?
• Which endpoint matters most to you next: video, better vision, or higher rate limits?

Links:

• Website: https://nflashx.cc.cd
• Docs: https://nflashx.cc.cd/docs.html
• GitHub: https://github.com/nflashx/nflashx