Getting started

This guide takes you from a fresh account to a rendered, on-brand design in a few minutes. The API is plain JSON over HTTP with a bearer token, so it works from any language — examples below are in TypeScript, Python, and cURL.

1. Create an API key

Log in to your Rivermark API dashboard and open Settings → API keys. Click Generate key, give it a name (for example, production or staging), and copy the key.

Set it as an environment variable so the examples below work as-is:

export RIVERMARK_API_KEY="rmk_live_…"

All requests go to https://api.rivermark.ai and carry the key as a bearer token:

Authorization: Bearer $RIVERMARK_API_KEY

2. Create your first design

This request creates an Instagram post from scratch, with no brand references, and prints the rendered output URL.

const res = await fetch("https://api.rivermark.ai/design", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.RIVERMARK_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    task: "Create an Instagram post announcing a summer sale: 30% off everything, this weekend only. Vibrant, sunny, confident.",
    output: [{ id: "post", size: "1080x1350" }],
  }),
});

const design = await res.json();
console.log(design.design_id);
// rmd_cj4hbqz2w0003mz6x7p9q2xe
console.log(design.output[0].url);
// https://cdn.rivermark.ai/rmd_cj4hbqz2w0003mz6x7p9q2xe/k3f8qz1w/post.png

A successful response looks like this:

{
  "design_id": "rmd_cj4hbqz2w0003mz6x7p9q2xe",
  "output": [
    {
      "id": "post",
      "url": "https://cdn.rivermark.ai/rmd_cj4hbqz2w0003mz6x7p9q2xe/k3f8qz1w/post.png",
      "format": "png",
      "width": 1080,
      "height": 1350
    }
  ],
  "change_summary": "Created an Instagram post for the summer sale",
  "change_description": "Built a full-bleed post with a sunny gradient, bold headline set in a geometric sans, and a sale badge in the lower third. Verified legibility and safe zones.",
  "changes": [
    {
      "kind": "created",
      "path": "frames/post.html",
      "url": "https://cdn.rivermark.ai/rmd_cj4hbqz2w0003mz6x7p9q2xe/k3f8qz1w/source/post.html"
    }
  ],
  "cost": 0.10
}

Open the output[0].url — that's your design, rendered exactly as it will appear.

3. Download the output

import { writeFile } from "node:fs/promises";

const res = await fetch(design.output[0].url);
await writeFile("post.png", Buffer.from(await res.arrayBuffer()));

4. Reuse or iterate on it (optional)

The design_id is persistent, and a design you've already made is the most useful reference you can give the next request. How you reuse it depends on what you want:

• As a template — pass it as a reference and ask for new content in its style: "Create a post from this template with the following copy: …" This is the basis of the user-created templates guide.
• As a style reference — pass it alongside other references and ask for something new that follows its look: "Create a story in the same style as this post."
• To iterate on it — pass it as continue_from to issue a follow-up request in the design's context. The original is effectively forked, and each turn returns a new design ID. This is how conversational editing works — see the AI design editor guide.

// Iterating: continue from the existing design
const next = await fetch("https://api.rivermark.ai/design", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.RIVERMARK_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    task: "Make the headline smaller and move the badge to the top right.",
    continue_from: design.design_id,
  }),
}).then((r) => r.json());

Next steps