CosyUI resolves your node graph automatically. CosyTemplates give you 150+ verified workflows ready to execute instantly. No Python. No broken nodes. No internet.
CosyUI is the workflow engine inside Promptus that turns any ComfyUI JSON file into a running workflow — without opening a terminal. Drag your .json file in, CosyUI reads the node graph, detects what's missing, installs dependencies with a single click, then runs your workflow locally on your GPU. Your existing Civitai folder, community shares, and custom setups all work without modification.
CosyTemplates are 150+ production-verified AI workflows built for image generation, video, 3D, audio, and API export. Every template is pre-tested and runs on first click — no hunting for missing models, no broken node errors, no configuration. Pick one, click run. Image to video. Text to 3D. Portrait to animation. Flux. Wan 2.1. VACE. Hunyuan. All available without a cloud account or per-generation cost.
Drag JSON in. Watch it resolve. See it generate. No Python, no terminal — this is the entire workflow.
Drag any ComfyUI workflow file into Promptus — yours, downloaded from Civitai, or shared by the community. CosyUI reads the node graph immediately.
Any missing nodes or models are flagged and installed automatically with one click. No manual pip installs. No hunting GitHub for the right version.
Click Run. Your workflow executes on your local GPU. When it works, save it as a CosyFlow — a reusable, one-click tool you can re-run, share, or export as a local API endpoint.
Install Python 3.10 exactly. Clone the repo. Install pip dependencies manually. Hunt GitHub for missing custom nodes. Fix version conflicts. Debug broken models. Typical time to first output: 2–4 hours.
Download Promptus. Drag your JSON. Click Install for missing nodes. Click Run. Typical time to first output: under 3 minutes. Zero terminal commands. Zero Python conflicts.
CosyTemplates are production-ready workflows curated from the best community setups and built internally by the Promptus team. Every template is tested, every model is confirmed available, every node is compatible. Pick one and generate — no configuration.
Drop any ComfyUI workflow. CosyUI maps the graph, flags what's missing, installs with one click, then runs. Your full archive, active again.
CosyTemplates are production-ready video, image, 3D, audio, and API workflows. No node-building. No configuration. Pick one, click run.
Flux, Wan 2.1, VACE, Hunyuan, Llama, Qwen and more — on your GPU. No API key, no cloud account, no per-generation cost.
Export any saved workflow as a callable local API endpoint. Use it in your product, automation, or client work. One click. Your AI backend, built.
Works with no internet. Not cached, not logged, not used for training. No cloud outage can touch your work. No server owns your output.
One payment of $49, and the app is yours permanently. Includes 12 months of model and workflow updates. RunwayML costs $1,140/yr.
| Feature | Promptus · CosyUI | Raw ComfyUI | RunwayML | Pika Labs |
|---|---|---|---|---|
| Import existing workflows | ✓ Drag, drop, run | ✓ Manual | ✗ | ✗ |
| No terminal / Python needed | ✓ Zero commands | ✗ Required | ✓ | ✓ |
| 150+ ready-to-run templates | ✓ CosyTemplates | ✗ Build your own | ✗ | ✗ |
| Auto-install missing nodes | ✓ One click | ✗ Manual pip | N/A | N/A |
| Runs fully offline | ✓ No internet | ✓ | ✗ Cloud only | ✗ Cloud only |
| Export workflow as local API | ✓ One click | ✗ | ✗ | ✗ |
| Data privacy | ✓ Never leaves device | ✓ | ✗ Cloud processed | ✗ Cloud processed |
| Annual cost | $49 once. Done. | Free + GPU costs | $1,140 / yr | $840 / yr |
Competitor pricing as published on RunwayML and Pika Labs list prices, checked September 2026. Plans and prices change — check their sites for current rates.
"The ComfyUI JSON import alone is worth $49. I dragged in my entire Civitai folder and now my workflows are available locally and API production-ready."
"I was paying $95/month for RunwayML. This does more, runs locally, costs $49 once. I did the maths in ten seconds. No regrets."
"Offline at midnight. Cloud was down. Promptus ran. Deadline met. That's what owning your tools actually feels like."
"The local API export turned my workflow into a product in one afternoon. I haven't looked back."
"I had 20 ComfyUI workflows sitting unused. I dragged them in, they all ran. I haven't opened a terminal since."
"Finally own my tools instead of renting them. Should have done this months ago."
Quotes reproduced with permission from public posts and beta feedback. Edited only for length.
ComfyUI is the open-source node-based AI workflow engine. CosyUI is an infrastructure engine inside Promptus with one-click dependency installation and a drag-and-drop experience. You get the power of ComfyUI without the terminal work — ComfyUI, made usable by non-developers.
Yes. CosyUI accepts any valid ComfyUI JSON workflow file. Drag it in, and CosyUI parses the node graph and tells you exactly what's needed. If nodes or models are missing, it installs them with one click. Workflows from Civitai, GitHub, Discord, or your own archive all work without modification.
CosyTemplates are verified workflows that ship with Promptus — tested by the team and ready to run on first click. CosyFlows are workflows you save yourself: once a workflow runs successfully in CosyUI, you can save it as a CosyFlow, making it a reusable click-to-run tool that can also be exported as a local API endpoint.
Yes. Once Promptus and your models are downloaded, no internet connection is required to run any workflow on the desktop app. Your prompts, inputs, and outputs never leave your machine — they are not logged, cached, or used for training. Cloud outages cannot interrupt your work. (Promptus Web, the browser version, does require a connection.)
Yes. Any saved CosyFlow can be exported as a callable local API endpoint with one click. This lets you integrate your AI workflow directly into your product, automation pipeline, or client project — without sending data to any cloud server. Your local GPU becomes your AI backend.
Promptus runs on Windows and macOS as a desktop app, and in the browser via Promptus Web. On Windows, local generation needs an NVIDIA GPU with at least 6GB VRAM for image workflows, and 12GB+ VRAM is recommended for video models like Wan 2.1. CosyTemplates are tagged with GPU requirements so you always know before you run.
Yes. Pay $49 once and own Promptus permanently — no monthly fees, no annual renewal, no subscription. That payment includes 12 months of model updates, new CosyTemplates, and CosyUI improvements. After 12 months the app keeps working exactly as it is; only new updates require a renewal. Compare that with RunwayML at $1,140/year or Pika at $840/year.
Own your AI tools instead of renting them.
150+ CosyTemplates. 100+ models. Fully offline. Yours forever.
$49 one-time payment · No subscription · 12 months of model updates included · Windows, macOS & browser · Desktop app runs 100% offline