ComfyUI Agnes-AI

English | 中文说明

ComfyUI custom nodes for the Agnes AI API — a free, cloud-based AI generation platform. Generate images, create videos, enhance prompts, and analyze visuals — all without a local GPU. Zero extra Python dependencies.

• 2026/09/17: Update ComfyUI-Agnes-AI to v1.2.1 ( updates.md )
  • Enabled HTTPS certificate verification by default; fixed API key management in Settings Panel.
• 2026/09/15: Update ComfyUI-Agnes-AI to v1.2.0 ( updates.md )
  • Upgraded to next-gen Agnes AI models: agnes-3.0-flash, agnes-image-2.5-flash, and agnes-video-2.5-flash.
  • Dynamic Input Slots: Nodes start clean with a single input point (image_0, ref_image_0, ref_audio_0) and automatically add new slots as you connect wires.
  • Zero Required Ports: All media connection points are fully optional; text-to-image and text-to-video work with prompt alone.
  • Comprehensive Video Resolutions: Added 480P, 720P, 1080P, 1K, and 2K options with automatic Flash 720P safeguard.
  • Reference Video Support: Added dedicated ref_video input for agnes-video-2.5 standard model.
  • Native Progress & Interruption: Integrated ComfyUI real-time progress bar and instant cancellation.
• 2026/07/29: Update ComfyUI-Agnes-AI to v1.1.0 ( updates.md )

Initial release with Agnes-AI Image, Video, Text, and Config nodes.

Get a free API key: platform.agnes-ai.com

• Image Generation (agnes-image-2.5-flash) — Text2img / img2img with ComfyUI 3.0 Autogrow reference images (starts with image_0, up to 4 images). Auto-detects mode based on inputs. 1K / 2K / 3K / 4K resolution.
• Video Generation (agnes-video-2.5-flash / agnes-video-2.5) — Text-to-video, image-to-video, first-and-last-frame animation, and multimodal reference (up to 5 images, 3 audios, 1 video). 480P–2K resolution tiers with Flash auto-safeguard, real-time progress bar, and frame/audio extraction.
• Prompt Enhancement (agnes-3.0-flash) — 4 built-in presets: enhance, translate, extract art style, describe image. 512K context with multimodal support.
• Settings Panel — Configure API key, select default models, all from ComfyUI's built-in Settings Panel. Supports multiple backup keys with automatic failover.

Search Agnes-AI in ComfyUI-Manager and install.

cd ComfyUI/custom_nodes
git clone https://github.com/1038lab/ComfyUI-Agnes-AI

Restart ComfyUI.

Download the latest release, extract to ComfyUI/custom_nodes/ComfyUI-Agnes-AI/, restart ComfyUI.

No pip install needed — zero additional dependencies.

Open ComfyUI Settings (⚙️ gear icon) → search "Agnes-AI" → enter your API key.

• Supports multiple backup keys (one per line or comma-separated) with automatic failover
• Alternatively, set the AGNES_API_KEY environment variable
• Key is saved to agnes_config.json and persists across restarts

Priority: env var (AGNES_API_KEY) > saved config (Settings Panel).

Generate images from text, or compose new images from up to 4 reference images with agnes-image-2.5-flash.

• Starts cleanly showing only image_0; automatically adds next slots as images are connected.
• Auto-detects text2img (no images connected) vs img2img (images connected).
• When images are connected without a text prompt, automatically defaults to merging them into a cohesive composition.
• Supports 1K / 2K / 3K / 4K with configurable aspect ratios. All connection points are optional.

Powered by agnes-video-2.5-flash (and agnes-video-2.5 / agnes-video-v2.0). Four generation modes:

• Text To Video — generate video purely from prompt (no inputs required)
• Image To Video — animate starting from first_image
• First and Last frame — interpolate smoothly between first_image and end_frame
• Reference (Images/Audio/Video) — generate video using multi-image, audio, and video reference materials

Inputs (All Connection Points Optional):

• mode — Text To Video, Image To Video, First and Last frame, Reference (Images/Audio/Video).
• prompt — Description of the video to generate.
• quality — 480P, 720P, 1080P, 1K, 2K. (Note: Flash model automatically downgrades to 720P per official specification).
• aspect_ratio — auto, 16:9, 9:16, 1:1, 4:3, 3:4, 21:9.
• duration (4–12s), seed.
• first_image — Start frame or primary reference image.
• end_frame — End frame for keyframe mode.
• ref_image_0 — Reference image slot (Autogrow: starts with ref_image_0, expands up to 5 images as connected).
• ref_audio_0 — Reference audio track slot (Autogrow: starts with ref_audio_0, expands up to 3 tracks as connected).
• ref_video — Reference video input (supported on standard agnes-video-2.5).

Outputs:

• video — Generated video file path.
• last_frame — Automatically extracted last frame of the video (IMAGE).
• frames — Full sequence of extracted video frames as a batched IMAGE tensor (requires local ffmpeg).
• audio — Extracted audio track as a standard ComfyUI AUDIO waveform (requires local ffmpeg).

Powered by agnes-3.0-flash. Process prompts through 8 built-in production presets:

Extensible Presets: All presets are externalized in presets/styles.json. You can customize them or drop any .json / .md preset files into the presets/ folder for automatic loading. Custom system prompt overrides are also supported on the node.

Ready-to-use workflows are available in the example_workflows directory:

• 01_Text_to_Image_and_Fusion.json — Text-to-image generation and multi-image fusion with Agnes-AI Image.

• 02_Text_and_Image_to_Video.json — Text-to-video and start-frame keyframe animation with Agnes-AI Video.

• 03_Prompt_Enhance_and_Vision.json — Prompt expansion and image reverse-prompt description with Agnes-AI Text.

Simply drag and drop any of these .json files into ComfyUI to get started immediately!

• Agnes AI — Free API and model infrastructure. Get your key at platform.agnes-ai.com
• Created by AILab

GPL-3.0

If this custom node helps you, please ⭐ the repo!