surface
kaitoi VS-1 compare
01 cmp comparison
All comparisons

Kaitoi and FLORA compared

For designers and creative teams who like FLORA's canvas and agent but want their own code in the graph, their own machines in the loop, or to pay as they go without a subscription.

FLORA is a hosted AI canvas for creative teams. Image, video, audio, text and 3D nodes connect on an infinite canvas, its agent FAUNA builds and runs workflows from a conversation, and Techniques package a workflow so anyone can run it as one node.

Kaitoi is also a hosted canvas of nodes with an agent, a timeline and apps for other people. The difference is underneath: every step is a node you can open, Penny writes new ones in Python, and your own computer or a GPU machine elsewhere can join the graph through Kaitoi CLI.

Checked against each product's own documentation on .

02 str strengths

What FLORA does well

Many models on one canvas. FLORA offers image, video, audio, text and 3D models from many providers in one account, with a published price for each model.

Workflows anyone can run. Techniques package a multi-step workflow as one node, the Technique Builder turns your own canvas into one, App Mode runs it without the canvas, and a reviewed community library shares them.

An agent that asks first. FAUNA builds, edits and runs workflows, and by default asks for approval with an estimated cost before it generates anything. Chatting with FAUNA is free on every plan.

Open to outside agents. A hosted MCP server lets Claude, ChatGPT, Codex, Cursor and others generate, edit canvases and run Techniques, FLORA publishes 57 agent skills, and an agent can build small FLORA Apps on your workspace.

Made for teams. Several people work on the same canvas at once, comments reach outside reviewers, and admins pool a budget and cap each member.

Finishing in the same place: a multi-track timeline that exports EDL, XML and FCPXML, a layer-based image editor, decks, and 3D export presets for common 3D tools.

03 apr approach

How Kaitoi approaches it

Many models in one account. More than 1,400 generative models and around 400 language models run on Kaitoi credits at the provider's own price with no markup, or, for providers that accept them, on your own keys. Studio shows a credit estimate before a node or graph runs.

Workflows anyone can run. Create subgraph folds a multi-step selection into one reusable node, a saved template starts new projects, nodes and templates publish to Community, and a workflow becomes a Station, a simple form or a custom app where each visitor runs on their own account, or an API endpoint.

An agent that shows its work. Penny builds, edits and runs graphs, writes the nodes that are missing, and edits storyboards and timelines. Each reply lists the tools it used, a stop icon ends the request, Studio shows a credit estimate before a run, and Penny gives one when asked. Penny has no approval step before generating, and its own model use is paid with credits.

Open to outside agents. Kaitoi's MCP server, on every plan, lets compatible AI clients build graphs, run approved work and inspect results, and on Studio write nodes. Penny calls outside MCP servers, and Kaitoi Code, the terminal agent in Kaitoi CLI, uses Kaitoi workflows from a code project. Penny can also write a node with its own HTML, CSS and JavaScript interface, and a Station can open with it as an app, invite only because it runs your own code.

Teams, without live co-editing. On a team Studio subscription, team workspaces have roles, shared projects, pooled credits and weekly caps per member. Share links, on every plan, are read-only. Edits do not sync live, and there are no comments on the canvas.

Finishing in the same project. A storyboard feeds a multi-track timeline that renders MP4, WebM or MOV in ProRes and exchanges OpenTimelineIO with editors, image edits run as nodes, and 3D exports as Scene GLB or renders in a connected Blender. OpenTimelineIO is the only interchange format, and there is no layer-based image editor.

Beyond that, every step is a node with a declared interface: a hosted model, a media tool, or Python. When the step you need does not exist, Penny writes it, and a node can carry its own HTML, CSS and JavaScript interface, WebAssembly included.

A run reuses the results of steps that did not change, so only the changed steps run and cost again.

And the graph is not limited to the cloud. Kaitoi CLI connects your own computer or a GPU machine elsewhere, so ComfyUI, Ollama, Blender and Runtime recipes run as nodes, and Kaitoi Code writes a recipe from a Hugging Face model or a GitHub repository.

04 tbl side by side

FLORA and Kaitoi, side by side

Checked November 13, 2026.
QuestionFLORAKaitoi
Models Which generative models you can reach.Image, video, audio, text and 3D models from many providers in one account, each with a published price. The Free plan covers text and image models.More than 1,400 generative models and around 400 language models from several providers as of September 2026, at the provider's own price with no markup, on credits or your own keys. When a provider releases a model that has no node yet, Penny can make one from the New Models panel. The catalogue is the same on every plan, with no subscription.
Agent An assistant that builds and runs the work with you.FAUNA builds, edits and runs workflows within the open project, and asks before generating, with an estimated cost, unless you switch it to Auto. Chatting with it is free.Penny builds and runs graphs, writes nodes, and edits storyboards, timelines and Lab boards. Each reply lists the tools it used, Studio shows a credit estimate before a run, and Penny gives one when asked, but there is no approval step before generating. Penny's own model use is paid with credits.
Your own steps and code Adding a step the product does not ship.Custom Actions turn a plain-language description into a tool on the canvas, and can hold Python or JavaScript that you or an agent write. We found no SDK for new node types.Every node is Python with a declared interface and packages from an approved list, and can carry its own HTML, CSS, JavaScript and WebAssembly interface. Penny writes nodes on every plan, editing code by hand needs Studio or Enterprise, and nodes can be published to Community.
Reuse and apps Packaging a workflow so you or others can run it again.Techniques run a whole workflow as one node, App Mode shares one as a simple app, an agent can build small FLORA Apps, and Techniques run through the API with webhooks.Create subgraph folds a workflow into one reusable node, and templates and Community share graphs and nodes. A Station turns a workflow into a form or a custom app where each visitor runs on their own account, Penny can write the node interface such an app opens with, a saved template becomes an API endpoint, and AI clients can run your workflows through Kaitoi's MCP server. Endpoint results are polled or streamed, with no webhooks.
Editing and finishing Assembling a cut, editing images, and handing off.On paid plans, a timeline of up to 10 tracks renders 30 fps MP4, with length caps by plan, and exports EDL, FCP7 XML or FCPXML. The Image Editor composites images, text, shapes, vectors and 3D as layers.A storyboard feeds a multi-track timeline that renders MP4, WebM or MOV in ProRes and exports and imports OpenTimelineIO, the only interchange format. Image edits run as nodes: inpaint, outpaint, erase, background removal, layer splitting and upscaling, with color nodes for levels, color wheels and LUTs. There is no layer-based image editor.
3D, vectors and sound Media beyond images and video.3D generation with export to GLB, glTF, FBX, OBJ, STL, PLY or USDZ, vector generation with SVG export, and speech, sound effects and music. We found no Gaussian splat support.3D generation through Tripo, Meshy, Hunyuan3D and World Labs with Scene GLB export, a Gaussian splat viewer and 3D Editor node, rendering in your own Blender, SVG from images or prompts, and ElevenLabs speech, sound effects and music.
Teamwork Working on the same project with other people.Real-time co-editing, comments that outside reviewers can answer, roles, share links with view or edit access, and a pooled budget with caps per member.Team workspaces with roles, shared projects and pooled credits on a team Studio subscription, and weekly caps per member. Edits do not sync live: when a teammate saves, Studio offers to load the newer version. Share links are read-only, and there are no comments on the canvas.
Running it again What happens to finished steps when you run a workflow again.Up to 50 nodes run at once in dependency order. Each run makes a new generation, and results are not cached.Independent nodes run at the same time, and a run reuses the results of steps that did not change, so only the changed steps run and cost again.
Your own hardware Running models on a machine you control.FLORA runs in the browser as a hosted service. We found no self-hosted or local option.Hosted, and your own computer or a GPU machine elsewhere can join through Kaitoi CLI: ComfyUI, Ollama, Blender, and Runtime recipes that Kaitoi Code writes from a Hugging Face model or GitHub repository.
Outside agents and code Driving the product from other software.A hosted MCP server, a REST API from the Starter plan, TypeScript and Go SDKs, a CLI and signed webhooks.Kaitoi's MCP server, on every plan, lets compatible AI clients build graphs, run approved work and inspect results, and on Studio write nodes. A REST API runs graphs from your own code, with no published SDK; endpoint results are polled or streamed, with no webhooks; and Kaitoi CLI brings Kaitoi Code, its terminal coding agent.

Each cell comes from the product's own website, documentation or code repository. "Not publicly documented" means we found no official page that says so, not that the feature is missing. Plans and features change; check the official pages before you decide.

05 fit which to choose

Which one fits

Choose FLORA if
  • Your team works on the same canvas at the same time and reviews with comments.
  • You want 57 published agent skills, and agent-built apps that read your FLORA workspace data.
  • You want a layer-based image editor, decks and Shopify export next to generation.
  • You want chatting with the agent to be free on every plan.
Choose Kaitoi if
  • You want any step written in Python, by Penny, or by hand on Studio, with its own interface when it needs one.
  • You want your own machines in the graph: a GPU box, ComfyUI, Ollama, Blender, or a model from a GitHub repository.
  • You want unchanged steps reused rather than generated again.
  • You want more than 1,400 models with no subscription, at the provider's own price.
  • You want to publish a workflow where each visitor pays for their own runs.
06 dem example

A real example: from a conversation to a commercial

Kaitoi Studio film

Kaitoi Signals #1 - From Conversation to Commercial

A simple conversation evolves into a complete ad as Penny moves through concept exploration, visual development, storyboard, shot generation, timeline, and rough cut assembly inside Kaitoi.

This film, from our Kaitoi Signals series, follows one ad from the first message to a rough cut.

  1. A conversation with Penny explores the concept.
  2. Penny moves through visual development and a storyboard.
  3. Shots are generated from the storyboard.
  4. The shots are assembled on the timeline into a rough cut, inside Kaitoi.

Read more in the documentation:

07 prc pricing

Pricing and access

FLORA

Per-seat plans from Free to Max, and custom Enterprise. Each paid seat adds the same monthly amount to a pooled budget, metered at published per-model prices, and unused usage does not roll over. Free covers text and image models, with watermarked images.

Kaitoi

Free to start, with no subscription. Credits pay for hosted models, including Penny's, at the provider's own price with no markup, and for paid Community nodes; steps that call neither cost nothing. Your own provider keys work on every plan and are billed by your provider. Studio adds included credits, team seats and hand-edited code.

08 faq questions

Questions people ask

Does Kaitoi have an agent like FAUNA?

Yes. Penny builds and runs graphs from a conversation, writes new nodes in Python, and edits storyboards and timelines.

Can several people edit a Kaitoi project at once?

Not live. Teams share projects with roles and pooled credits, and when a teammate saves, Studio offers to load the newer version. FLORA supports real-time co-editing.

Can I reuse a workflow the way a Technique does?

Yes. Create subgraph folds a workflow into one reusable node, and you can save the graph as a template, publish nodes and templates to Community, or turn the workflow into a Station or an API endpoint.

Does Kaitoi have an MCP server like FLORA?

Yes, on every plan. Compatible AI clients can build graphs, run approved work and inspect results through it, and Penny can also call outside MCP servers.

09 mor more comparisons

Other comparisons