Templates

Templates hold the configuration an instance launches from. See Create a template for what the fields mean in practice.

List templates

GET /api/profile/templates Session or API key

Returns your templates and the container images you can choose from.

{
  "templates": [
    {
      "id": 87,
      "title": "ROCm PyTorch dev",
      "slug": "rocm-pytorch-dev",
      "image": "rocm/pytorch:latest",
      "instance_type": "jupyter",
      "ssh_enabled": true,
      "enabled": false,
      "sort_order": 0
    }
  ],
  "images": [
    { "name": "ROCm PyTorch", "image": "rocm/pytorch:latest", "enabled": true }
  ]
}

Create a template

POST /api/profile/templates Session or API key

Parameter	Type		Description
title	string	Required	Display name.
image	string	Required	Container image, from the catalog.
instance_type	string	Optional	jupyter, gradio, streamlit, comfyui, vllm, sglang, custom.
description	string	Optional	Longer description.
slug	string	Optional	URL-friendly identifier.
category	string	Optional	Grouping label.
tags	array	Optional	Free-form tags.
repo_url	string	Optional	Git repository cloned on start.
branch	string	Optional	Branch to check out. Defaults to main.
notebook_path	string	Optional	Notebook to open on start.
start_command	string	Optional	Command run at startup. Required for vllm and sglang.
app_port	integer	Optional	Port the app serves on. Must be one the platform routes: 7860 Gradio, 8501 Streamlit, 8188 ComfyUI, 8000 vLLM, 30000 SGLang.
ssh_enabled	boolean	Optional	Allow SSH into instances from this template. Defaults to false.
cover_url	string	Optional	Cover image URL.
sort_order	integer	Optional	Position in your list.
enabled	boolean	Optional	Publish publicly. Requires editor permission; forced to false otherwise.

curl -X POST https://radeon-global.anruicloud.com/api/profile/templates \
  -H "Authorization: Bearer $RADEON_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Qwen 7B on vLLM",
    "image": "rocm/vllm:latest",
    "instance_type": "vllm",
    "start_command": "vllm serve Qwen/Qwen2.5-7B-Instruct --host 0.0.0.0 --port 8000",
    "app_port": 8000
  }'

An app_port outside the routable set is rejected with 400.

Update a template

PUT /api/profile/templates/{template_id} Session or API key

Same body as creation. 404 if the template doesn’t exist or isn’t yours.

Changes apply to future launches. A running instance keeps the configuration it started with.

Delete a template

DELETE /api/profile/templates/{template_id} Session or API key

{ "success": true }

Deleting a template doesn’t affect an instance already running from it.

Launch a template

POST /api/templates/{template_id}/launch Session or API key

Parameter	Type		Description
gpu_count	integer	Optional	1, 2, or 4. Defaults to 1.
launch_verification_token	string	Optional	Required only when the account has launch verification enabled.

Returns the same status object as POST /api/notebook/request. Poll GET /api/notebook/status from there.

Launches are rate limited to roughly one per minute, three per ten minutes, and five per hour. Exceeding that returns 429 with a Retry-After header.

Browse public templates

GET /api/templates None; more with a credential

Returns published templates. Authenticated callers also see their own private ones. Useful for finding a starting point without creating anything.