---
name: nocam
description: Make AI photos, videos and text through the NoCam API. Use when the person asks to generate, edit or animate a photo, make a clip, or call a model such as GPT Image, Nano Banana, Seedance, Muse or Omni.
---

# NoCam API

One key for many AI photo, video and text models. Base URL: `https://api.nocamclub.com/v1`. The key is in `NOCAM_API_KEY`: never print it.

## Before you start

Read https://docs.nocamclub.com/llms.txt for the map, and https://docs.nocamclub.com/for-ai-agents.md for the rules. Read a page by adding `.md` to its address.

## How

1. `GET /v1/models` shows what's on sale. `GET /v1/models/{id}` gives `input_schema` and `prices`. Never guess a model's options.
2. Photos and videos of the person's: `POST /v1/uploads` with `{"url": …}` or `{"content_type": …, "bytes": …}`. In the prompt they are `img1`, `video1`, `audio1`.
3. `POST /v1/estimate` checks the request and gives the price, for free. If it's more than 100 credits ($1), ask first.
4. `POST /v1/generations` with an `Idempotency-Key`. Photos: add `Prefer: wait=60`. Videos: poll `GET /v1/generations/{id}` every `poll_after_ms`.
5. Download the result from `outputs[0].url`: links stop after 7 days.

## When it goes wrong

Read `error.code`, `error.hint` and `error.docs_url`. A failed generation gives its credits back by itself. Tell the person about refusals, and don't reword a prompt to get past them. Never top up on your own: send the person to `top_up_url`.

To test without spending, use a `nc_test_…` key: sample results and no credits.