Rodin AI API Guide

The Rodin API exposes the exact same Gen-2.5 generation pipeline that powers text-to-3D and image-to-3D in the browser, just reachable over HTTP instead of a UI. That makes it the right layer for anything that shouldn't be a person clicking "generate" one asset at a time: turning a spreadsheet of product names into a batch of models, wiring generation into a Blender, Unity or Unreal plugin, or running Rodin as one step inside a larger internal pipeline.

It's built for programmatic and batch use specifically - async job submission, free status polling, and a download step that returns URLs rather than streaming a file inline, so it fits naturally into a queue-based system rather than a single synchronous request.

API access is gated by plan, separately from what a plan allows in the web app. Free and Creator can generate through the browser but can't call the API - Business is the entry point for programmatic access.

Exact requests-per-minute ceilings for Business aren't published in a hard SLA the way Enterprise's are - treat 120-240 RPM as the working range rather than a contractual number. Full plan pricing is on the pricing page.

Parameters on POST /rodin fall into four groups. Nothing here needs to be set for a working call - a bare prompt or a single image is enough - but each group gives you the same controls available in the web app.

Input

• images - up to 5 files for image-to-3D, combinable with a prompt
• prompt - required for text-to-3D, optional alongside images
• image_label - directional hints per image (F/FL/FR/L/R/B/BL/BR/U/D/?)
• use_original_alpha - boolean, preserves source transparency

Generation config

• tier - Extreme-Low, Low, Medium, High or Extreme-High effort
• quality - high / medium / low / extra-low
• mesh_mode - Raw (triangular) or Quad topology
• geometry_file_format - glb / usdz / fbx / obj / stl
• geometry_instruct_mode - faithful or creative interpretation of the prompt
• is_micro - micro-detail geometry, Extreme-High tier only

Materials & texture

• material - PBR / Shaded / All / None
• texture_mode - legacy / extreme-low / low / medium / high
• hd_texture - boolean, higher-resolution textures
• texture_delight - removes baked-in lighting from the texture

Advanced

• seed - 0-65535, for reproducible results
• TAPose - boolean, outputs a rigging-friendly T/A pose
• bbox_condition - array [Width, Height, Length] to constrain proportions
• preview_render - boolean, returns a quick preview render
• addons - array, e.g. HighPack for 4K textures (+1 credit)

Each example submits a prompt-only job and polls until it's done - the shortest path to a first result. Swap in your own API key and prompt to run it.

# 1. submit the job
curl -X POST https://api.hyper3d.com/api/v2/rodin \
  -H "Authorization: Bearer $RODIN_API_KEY" \
  -F "prompt=a weathered brass compass, game-ready prop" \
  -F "tier=Medium"

# 2. poll until status is complete
curl https://api.hyper3d.com/api/v2/status/<task_uuid> \
  -H "Authorization: Bearer $RODIN_API_KEY"

import requests, time

headers = {"Authorization": f"Bearer {API_KEY}"}

resp = requests.post(
    "https://api.hyper3d.com/api/v2/rodin",
    headers=headers,
    data={"prompt": "a weathered brass compass, game-ready prop", "tier": "Medium"},
)
task_id = resp.json()["uuid"]

while True:
    status = requests.get(f"https://api.hyper3d.com/api/v2/status/{task_id}", headers=headers).json()
    if status["status"] == "complete":
        break
    time.sleep(3)

files = requests.post(
    "https://api.hyper3d.com/api/v2/download",
    headers=headers, json={"task_uuid": task_id},
).json()

const form = new FormData();
form.append("prompt", "a weathered brass compass, game-ready prop");
form.append("tier", "Medium");

const submit = await fetch("https://api.hyper3d.com/api/v2/rodin", {
  method: "POST",
  headers: { Authorization: `Bearer ${API_KEY}` },
  body: form,
});
const { uuid } = await submit.json();

// poll /status/{uuid} until status === "complete", then:
const files = await fetch("https://api.hyper3d.com/api/v2/download", {
  method: "POST",
  headers: { Authorization: `Bearer ${API_KEY}`, "Content-Type": "application/json" },
  body: JSON.stringify({ task_uuid: uuid }),
}).then(r => r.json());

The API draws from the same credit pool as your web account - there's no separate API balance to manage. Credits are spent on generation only; checking status and downloading are always free, so you can poll as often as you like without worrying about cost.

Direct credits beyond your plan allowance are $1.50 each and don't expire. Full breakdown of plans and included credits is on the pricing page.

Errors return a code you can branch on programmatically rather than just an HTTP status:

SUBSCRIPTION_PLAN_TOO_LOW is the one you'll hit if you try to call the API from a Free or Creator account - upgrade to Business to clear it. INSUFFICIENT_FUND means your credit balance is short; top up or wait for the next billing cycle.