ComfyUI-fal-API
Production-ready ComfyUI custom nodes for 1,400+ fal.ai models, auto-updated image, video, audio, LLM and VLM APIs with native media, caching and cost controls.
What is ComfyUI-fal-API?
Production-ready ComfyUI custom nodes for 1,400+ fal.ai models, auto-updated image, video, audio, LLM and VLM APIs with native media, caching and cost controls.
Overview of ComfyUI-fal-API
What it is
ComfyUI-fal-API is a set of custom nodes for ComfyUI that integrates the entire fal.ai model catalog into the ComfyUI workflow. It includes approximately 90 hand-written nodes for popular models and about 1,400 auto-generated nodes covering every live public model on fal, including image, video, audio, 3D, LLMs, and more. A single FAL_KEY unlocks all nodes. The pack features a persistent result cache (identical calls are served from disk for free), spend guards, async execution, URL passthrough for zero-I/O chaining, and platform utilities for parallel fan-out and cost estimation. The exact current catalog is listed in MODELS.md.
Key features
- Async execution on modern ComfyUI keeps the UI responsive while jobs run.
- Persistent result cache: identical fal calls are served from disk, free and instant, across restarts; uploads are deduplicated.
- Spend guard: refuse to submit jobs past a session budget or below a balance floor.
- URL passthrough: wire a fal node's URL output into the next node's direct URL input to avoid intermediate media transfer.
- Auto-generated nodes for all live fal models, with native IMAGE/VIDEO/AUDIO sockets, schema-derived tooltips, and pricing in the help panel.
- Platform utilities: Fal Submit + Fal Collect for parallel fan-out, Fal Result by Request ID to re-fetch past results without re-paying, Fal Cost Estimator, and Fal Session Costs.
- Editable suggestions: modes, voices, languages, and model IDs with short example values get dropdowns with an “Enter custom value…” option.
Use cases
- Running any fal.ai model (image generation, video, audio, 3D, LLMs) directly inside ComfyUI workflows.
- Building complex pipelines that chain multiple fal models with zero-I/O intermediate media.
- Using the persistent cache to avoid paying for repeated identical calls.
- Parallel execution of multiple jobs (e.g., five video jobs concurrently) via Fal Submit/Collect.
- Dataset preparation: images to captioned training ZIP, media loading from URLs/folders, video trim/concat/mux/frame-extract, image grids and resizing.
- Re-fetching past results by request ID without additional cost.
Getting started
Install via ComfyUI Manager (search for ComfyUI-fal-API) or manually: cd custom_nodes git clone https://github.com/gokayfem/ComfyUI-fal-API.git cd ComfyUI-fal-API python -m pip install -r requirements.txt For ComfyUI Windows Portable, use its embedded interpreter: .python_embededpython.exe -m pip install -r .ComfyUIcustom_nodesComfyUI-fal-APIrequirements.txt Configure your FAL_KEY by copying config.ini.example to config.ini and replacing the placeholder, or set the FAL_KEY environment variable. Restart ComfyUI. Curated nodes appear under the FAL category, auto-generated nodes under FAL/Models/, and hand-picked models under FAL/Featured.
License and pricing
Licensed under Apache-2.0. The project is open source; no pricing information is provided beyond the license.
Based on the project's public README and repository metadata.