CNAPS is a no-code builder that chains 80+ image, video, audio and text AI models into one workflow. Free plan; paid plans from $20/month.
Researched and fact-checked by Bartu Cavusoglu · Updated October 26, 2026
CNAPS is a no-code visual builder that connects image, video, audio and text AI models into a single workflow. Instead of asking one general-purpose model to do every job, users chain specialized models so each step runs on the model built for that task. The site lists more than 80 models that can be connected, and any of them can be swapped at any step.
The company describes itself as a no-code visual AI startup founded by former Google and NVIDIA engineers. Its first product, CNAPS Studio, is a drag-and-drop builder for enhancing images and videos while also integrating text and speech. Today the same workflows also run from ready-made recipes and from outside tools through an API and MCP.
CNAPS offers two ways to build. In the pick-and-go mode, a user chooses what they want to make and the right models chain themselves into a result, with no nodes and no setup. In CNAPS Studio, users connect models themselves on a canvas and build a custom pipeline, swapping a node whenever they want to try a different model.
Workflows accept one input or many, in the form of text, image, video or audio, and can return image, video, audio or text. Each step's output becomes the next step's input, and the interface shows what changed at each step. Every run consumes GPU time, which depends on the model, resolution and settings.
The model library covers video generation (including Sora 2 and MiniMax-H3), image generation (Nano Banana, GPT Image, FLUX.2), upscaling (SeedVR2 7B, SwinIR), image editing, background removal, sound and music, and reasoning models such as Claude, ChatGPT and Gemini. Users can create unlimited flows on any plan and publish them to a Community where others can fork them.
For volume work, Batch Run sends many inputs, such as a set of product photos, through one flow. Pro adds parallel GPU processing so queued runs finish sooner. Every plan includes API keys for tools such as n8n and Zapier, and MCP lets assistants such as Claude find models, build a flow and run it from a chat.
CNAPS is a no-code builder that connects image, video, audio and text AI models into one workflow. Users can pick from 80+ models and swap them at any step. It runs from a visual canvas, ready-made recipes, an API and MCP.
CNAPS has a Free plan, Starter at $20/month, Pro at $75/month and Business at $200/month with 4 seats included. Extra Business seats cost $50 each. Enterprise pricing is custom, and yearly billing saves 17%.
Yes. Eligible workspaces can try Pro free for two weeks with no credit card required. If no payment method is added, the workspace stays on Free after the trial. Each workspace can use the trial once.
All of them, on every plan including Free. Plans differ in GPU time, queueing and processing speed, not in model access.
CNAPS can be used from automation tools such as n8n and Zapier through an API key. AI assistants such as Claude can connect through MCP to build and run flows from a chat. Both are available on every plan.
Yes. Outputs created on the Free plan carry a watermark, and every paid plan removes it.
Business offers a team workspace with roles, shared GPU time and flows. Enterprise adds dedicated SaaS, private cloud or air-gapped on-premises deployment, SSO via SAML or OIDC, SCIM and SLA-backed support.
No. CNAPS is built for no-code use, with a pick-and-go mode and a drag-and-drop canvas. Developers can also use the API and MCP if they want.