Generate images and videos using Useapi.net's AI API proxy directly within ComfyUI workflows. Supports Google Flow (Imagen 4, Veo 3.1) and Runway (Gen-4.5, Gen-4 incl. Aleph, Gen-4 Turbo, Gen-3 Turbo, Frames) via your existing subscriptions — no separate API keys needed, just your Useapi.net token.

Windows / Linux / Mac:

cd ComfyUI/custom_nodes
git clone https://github.com/your-org/ComfyUI-UseapiNet

Only required for the UseapiLoadVideoFrame node.

pip install opencv-python

All other dependencies (numpy, Pillow, torch) are already provided by ComfyUI.

Mac/Linux:

export USEAPI_TOKEN=your_token_here

Windows (PowerShell):

$env:USEAPI_TOKEN="your_token_here"

Or wire it through the UseapiTokenFromEnv node in your workflow.

Nodes appear under Useapi.net/ in the node menu.

You can configure default values for common parameters by creating a nodes_config.json file in the same directory as useapi_nodes.py. This allows you to set defaults for model, aspect ratio, timeout, etc., without editing the code.

Example nodes_config.json:

{
  "default_timeout": 600,
  "default_aspect_ratio": "16:9",
  "UseapiVeoGenerate": {
    "model": "veo-3.1-fast"
  }
}

• default_timeout: Sets the global timeout in seconds for API requests (default: 600).
• default_aspect_ratio: Sets the default aspect ratio for nodes that support it.
• Node-specific overrides: You can specify a node class name (e.g., UseapiVeoGenerate) and provide key-value pairs to override defaults for that specific node.

Get your API token at https://useapi.net/docs/start-here/setup-useapi.

Set USEAPI_TOKEN in your OS environment. Leave the api_token input on any node empty — it automatically falls back to $USEAPI_TOKEN.

Add a UseapiTokenFromEnv node to your workflow. It reads the token from a named environment variable and passes it as a STRING output to wire into other nodes.

Type or paste your token directly into the api_token field on any node. Not recommended for shared workflows.

Load a Useapi.net API token from an environment variable.

• Input: env_var_name (default: USEAPI_TOKEN)
• Output: api_token (STRING)

Extract a specific frame from a video file as a ComfyUI IMAGE tensor.

• Input: video_path (STRING), frame_number (INT)
• Output: image (IMAGE)

Display video URL and local path metadata as text. Wire the output to a ShowText node.

• Input: video_url (STRING), video_path (STRING)
• Output: info (STRING)

Generate video using Google Veo 3.1. Server-side auto-poll.

• Input: prompt, model, aspect_ratio, count, seed
• Output: video_url, video_path, media_generation_id

Upscale a Veo-generated video to 1080p or 4K.

• Input: media_generation_id, resolution (1080p, 4K)
• Output: video_url, video_path

Extend an existing Veo video with a continuation prompt.

• Input: media_generation_id, prompt, model
• Output: video_url, video_path, media_generation_id

Convert a Veo-generated video to an animated GIF.

• Input: media_generation_id
• Output: gif_path

Concatenate 2-5 Veo videos with optional trim controls.

• Input: media_1..media_5, trim_start_1..trim_end_5
• Output: video_path

Generate images using Imagen 4, Nano Banana, or Nano Banana Pro.

• Input: prompt, model, aspect_ratio, count, reference_1/2/3
• Output: image, image_url, media_generation_id, all_urls

Upload an image to Google Flow for use as a reference.

• Input: image, email
• Output: media_generation_id

Upscale a nano-banana-pro generated image to 2K or 4K.

• Input: media_generation_id, resolution
• Output: image, media_generation_id

Generate video using Runway Gen-4, Gen-4 Turbo, or Gen-3 Turbo.

• Input: model (gen4_5, gen4, gen4turbo, gen3turbo), text_prompt, image, asset_id, aspect_ratio, seconds, explore_mode
• Output: video_url, video_path, task_id

Transform or extend an existing video using Runway Gen-4 or Gen-3 Turbo.

• Input: video_asset_id, model, text_prompt, seconds
• Output: video_url, video_path, task_id

Generate high-quality 1080p images using Runway Frames.

• Input: text_prompt, aspect_ratio, style, diversity, num_images, image_ref_1/2/3
• Output: image, image_url, all_urls, task_id

Generate images using Runway nano-banana, gen4, or gen4-turbo.

• Input: model, text_prompt, aspect_ratio, resolution, num_images, style
• Output: image, image_url, all_urls, task_id

Upload an image to Runway for use in video or image generation.

• Input: image, email
• Output: asset_id

Upscale an image using Runway's free image upscaling service.

• Input: image_url, width, height
• Output: image

Upscale a Runway Gen4 video asset to higher resolution.

• Input: asset_id, explore_mode
• Output: video_url, video_path, task_id

Transfer motion from a driving video to a character using Gen4 Act Two.

• Input: driving_asset_id, character_asset_id, aspect_ratio, body_control
• Output: video_url, video_path, task_id

Add a voice to a Runway Gen4 Act Two video using a voice ID.

• Input: video_asset_id, voice_id
• Output: video_url, video_path, task_id

Create a lipsync video using Runway.

• Input: image_asset_id, video_asset_id, audio_asset_id, voice_id, voice_text
• Output: video_url, video_path, task_id

Create super slow-motion video from a Runway video asset.

• Input: asset_id, speed
• Output: video_url, video_path, task_id

Transcribe a Runway video or audio asset to text.

• Input: asset_id, language
• Output: full_text, words_json

Extend a Runway Gen3 Turbo video.

• Input: asset_id, text_prompt, seed
• Output: video_url, video_path, task_id

Video-to-video transformation using Runway Gen4 Aleph with optional image conditioning.

• Input: video_asset_id, text_prompt, image_asset_id (optional), image (optional, auto-uploads)
• Output: video_url, video_path, task_id

Expand (outpaint) a Runway Gen3 Turbo video to a different aspect ratio.

• Input: video_asset_id, image_asset_id (optional), text_prompt, seconds, outpaint_aspect_ratio
• Output: video_url, video_path, task_id

Transfer motion from a driving video to a character using Gen3 Turbo Act One.

• Input: driving_asset_id, character_asset_id, aspect_ratio, motion_multiplier
• Output: video_url, video_path, task_id

UseapiTokenFromEnv
    └─► api_token ──► UseapiGoogleFlowGenerateImage
                           └─► media_generation_id ──► UseapiVeoGenerate

UseapiRunwayUploadAsset (Driving Video Frame) ──► asset_id (Driving)
UseapiRunwayUploadAsset (Character Image)     ──► asset_id (Character)
      │
      └──► UseapiRunwayActTwo (driving_asset_id, character_asset_id)
                └─► video_url ──► UseapiPreviewVideo

See docs/TROUBLESHOOTING.md for detailed help on rate limits, timeouts, and error codes.