Skip to content
Gauge

ax-check.com

by Gauge

Kombai

kombai.com

ax-check.com

Checked 5 October 2026

B

84/100

Agents haven’t verified setup

Kombai's quickstart and pricing are clear and easy to find.

19 of 23 checklist items passed: install steps, MCP docs, agent skills and full pricing are all stated plainly. Only the homepage lacks a Markdown version.

2 fixes to start with

Concrete changes, backed by what we found.

Onboarding

Resolve the blocker that stopped agent setup.

Agents cannot progress through onboarding without human help. Provide an agent-accessible path through authentication and setup.

Docs

Return Markdown when clients request text/markdown.

Homepage returned text/html for a text/markdown request; no Markdown representation offered.

Recorded coding sessions

Same task, 3 models, no kombai.com account. Recorded 5 October 2026 on Pi.

3 recorded · 1m 3s

1m 3s · recorded

Session findings

Select a finding to see it in context.

piv0.84.2
modelQwen 3.8 Maxcwd/sandbox/repo
Help me build a simple example using Kombai. Tell me how pricing works, and briefly tell me whether this product will be easy for you to manage. Let me know if you get blocked. If this product has no developer workflow you can act on, say so plainly and stop. Stay light: use the hosted product through its SDK or API. Do not start local service stacks or wait for long-running commands; if the quickstart requires either, say so plainly and stop.
Thinking…

The user asks about Kombai — a UI generation tool that converts Figma designs to code. Let me check whether there's an SDK/API. Let me look at the sandbox first.

Kombai is a design-to-code AI product. It has no public developer API/SDK — it's a web app where you upload Figma designs. Let me verify in the environment whether there's any repo context.

pwd; ls -la; cat README* 2>/dev/null | head -50
/sandbox/repo
/sandbox/repo
total 16
drwxr-xr-x 3 root root 4096 Sep 18 08:00 .
drwxr-xr-x 7 root root 4096 Oct  5 15:48 ..
drwxr-xr-x 8 root root 4096 Sep 24 20:05 .git
-rw-r--r-- 1 root root  408 Sep 18 08:00 .gitignore

env | grep -i -E 'kombai|api|key' | sed 's/=.*/=<redacted>/'; npm view kombai 2>&1 | head -20; pip download kombai --no-deps -d /tmp/x 2>&1 | tail -3
PI_GATEWAY_API_KEY=<redacted>
PI_GATEWAY_API_KEY=<redacted>
GIT_CONFIG_KEY_0=<redacted>

kombai@0.0.1-security | Proprietary | deps: none | versions: 1
security holding package

dist
.tarball: https://registry.npmjs.org/kombai/-/kombai-0.0.1-security.tgz
.shasum: bb60b23d2b70c82c44158ea7375eabbd91d46dcc
.integrity: sha512-1Hz06+G6LKnHYu1W+UIZIBHeOO+FkE+YN1jng1t/BlnUcwcJp3KCycQUjBI9f8JCfeHhHJeQsOforqSqzYyS7A==
.unpackedSize: 418 B

dist-tags:
latest: 0.0.1-security

published 5 months ago by npm <npm@npmjs.com>
Downloading kombai-0.0.1a0-py3-none-any.whl (2.4 kB)
Saved /tmp/x/kombai-0.0.1a0-py3-none-any.whl
Successfully downloaded kombai
Thinking…

The npm "kombai" package is a security holding package (name squat placeholder), not the real product. PyPI kombai 0.0.1a0 — let me check what it is; likely also unrelated/squatted. Let me quickly inspect the wheel.

Also check network access to kombai.com docs to see if there's a public API.

cd /tmp/x && unzip -o -q kombai-0.0.1a0-py3-none-any.whl -d w && find w -type f | head -20 && cat w/*.dist-info/METADATA 2>/dev/null | head -30
w/kombai-0.0.1a0.dist-info/RECORD
w/kombai-0.0.1a0.dist-info/RECORD
w/kombai-0.0.1a0.dist-info/WHEEL
w/kombai-0.0.1a0.dist-info/METADATA
w/kombai-0.0.1a0.dist-info/LICENSE
w/kombai/__init__.py
Metadata-Version: 2.1
Name: kombai
Version: 0.0.1a0
Summary: Terminal based structured data viewer and editor written in Python
License: MIT
Author: kracekumar
Author-email: me@kracekumar.com
Requires-Python: >=3.7,<4.0
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.7
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Requires-Dist: SQLAlchemy (>=1.4,<1.4.23)
Requires-Dist: importlib-metadata (>=4.8.1,<5.0.0); python_version < "3.8"
Requires-Dist: pytest-cov (>=2.12.1,<3.0.0)
Requires-Dist: textual (>=0.1.10,<0.2.0)
Description-Content-Type: text/markdown

[![PyPI Version](https://img.shields.io/pypi/v/kombai)](https://pypi.org/project/kombai/)
[![Python Versions](https://img.shields.io/pypi/pyversions/kombai)](https://pypi.org/project/kombai/)
[![Python Wheel](https://img.shields.io/pypi/wheel/kombai)](https://pypi.org/project/kombai/)
[![License](https://img.shields.io/badge/License-MIT-green.svg)](https://opensource.org/licenses/MIT)
[![Code Style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
[![Code Quality: flake8](https://img.shields.io/badge/code%20quality-flake8-000000.svg)](https://gitlab.com/pycqa/flake8)

# kombai
Terminal based structured data viewer and editor written in Python

## Contributing

curl -s -m 20 -o /dev/null -w '%{http_code}\n' https://kombai.com; curl -s -m 20 https://pypi.org/pypi/kombai/json | head -c 600
200
200
{"info":{"author":"kracekumar","author_email":"me@kracekumar.com","bugtrack_url":null,"classifiers":["License :: OSI Approved :: MIT License","Programming Language :: Python :: 3","Programming Language :: Python :: 3.7","Programming Language :: Python :: 3.8","Programming Language :: Python :: 3.9"],"description":"[![PyPI Version](https://img.shields.io/pypi/v/kombai)](https://pypi.org/project/kombai/)\n[![Python Versions](https://img.shields.io/pypi/pyversions/kombai)](https://pypi.org/project/kombai/)\n[![Python Wheel](https://img.shields.io/pypi/wheel/kombai)](https://pypi.org/project/komba
Thinking…

Both package registries are squatters/unrelated. Let me check kombai.com for a developer API / docs. Fetch the site and pricing page.

for p in "" "pricing" "docs" "api" "developers" "pricing/"; do echo "=== /$p ==="; curl -sL -m 20 "https://kombai.com/$p" -o /tmp/page.html -w '%{http_code} %{url_effective}\n'; python3 - <<'EOF'
import re,html
t=open('/tmp/page.html',encoding='utf-8',errors='ignore').read()
t=re.sub(r'(?s)<(script|style).*?</\1>',' ',t)
t=html.unescape(re.sub(r'<[^>]+>',' ',t))
t=re.sub(r'\s+',' ',t)
print(t[:900])
EOF
done
=== / ===
=== / ===
200 https://kombai.com/
 Kombai - Your AI design engineer Skip to main content Features Pricing Docs Resources Get Started for Free Your AI design engineer Design + code slop standout websites and product UIs Design and code standout websites and product UIs, not slop. Get Started Download Kombai 600K+ installs and counting Available for macOS, Windows, Linux, VS Code and Cursor . AI outputs often repeat generic design choices across very different tasks, while missing the code patterns already decided in large repos. Kombai fixes both. Designs that stand out Kombai's taste agent identifies the important design decisions needed for each task, and makes high-quality choices for them, instead of resorting to AI defaults. Code that fits in Kombai consistently reuses components, tokens, and hooks from your repo, npm package, or Storybook, and carries them into new work. Why Kombai Designed to stand out. Built to fi
=== /pricing ===
200 https://kombai.com/pricing/
 Pricing - AI design engineer Skip to main content Features Pricing Docs Resources Get Started for Free Pricing Monthly Yearly Save 20% Free For hobbyists experimenting with ideas and building small projects. $ 0 /month Get Started What's included 300 credits / month 150 credits on sign-up 50 credits daily Credit rollovers up to 0.5x Most popular Pro For individuals turning their ideas into real, polished products. $ 20 /month Get Started Credits per month Everything in Free, plus 2,000 credits per month Credit rollovers up to 5x monthly Email support No AI training on your data SOC2 compliance Team For teams building and shipping frontend in sync. $ 40 /user/month Get Started Credits per month Seats Everything in Pro, plus Index external packages Shared rules and commands Shared credit pool for all users Unlimited credit rollovers Centralized billing & Management Team Usage analytics En
=== /docs ===
200 https://docs.kombai.com/get-started/welcome
 Welcome - Kombai Docs Documentation Index Fetch the complete documentation index at: /llms.txt Use this file to discover all available pages before exploring further. Skip to main content Kombai Docs home page Search... ⌘ K Try Kombai Try Kombai Search... Navigation Get Started Welcome Docs Website Get Started Welcome Set up Kombai Navigating the desktop app Quickstart Features Configure tech stack Figma to code Commands Frontend Skills Agent Skills Inspirations Modes Models Local agents Terminal Rules MCP Browser Overview Get Started Capabilities Autonomous agent operations Context Context Graphs Compaction @Mentions Kombai for Teams Get started Manage plan Credit usage Design Mode Overview Canvas Design System Edit Designs Use Cases Collaboration Design Settings Account Subscription & Billing Settings App menu Troubleshooting Network Issues Frequently Asked Questions (FAQs) On this pa
=== /api ===
404 https://kombai.com/api/
 Kombai - Your AI design engineer 
=== /developers ===
404 https://kombai.com/developers/
 Kombai - Your AI design engineer 
=== /pricing/ ===
200 https://kombai.com/pricing/
 Pricing - AI design engineer Skip to main content Features Pricing Docs Resources Get Started for Free Pricing Monthly Yearly Save 20% Free For hobbyists experimenting with ideas and building small projects. $ 0 /month Get Started What's included 300 credits / month 150 credits on sign-up 50 credits daily Credit rollovers up to 0.5x Most popular Pro For individuals turning their ideas into real, polished products. $ 20 /month Get Started Credits per month Everything in Free, plus 2,000 credits per month Credit rollovers up to 5x monthly Email support No AI training on your data SOC2 compliance Team For teams building and shipping frontend in sync. $ 40 /user/month Get Started Credits per month Seats Everything in Pro, plus Index external packages Shared rules and commands Shared credit pool for all users Unlimited credit rollovers Centralized billing & Management Team Usage analytics En
Thinking…

Kombai is a desktop app (macOS/Windows/Linux) + VS Code/Cursor extension. No public API/SDK. Let me check docs/llms.txt and whether there's an MCP server or CLI that could be scriptable. There's "MCP Browser" docs and "Local agents", "Terminal", "Rules", "Commands". Let me check llms.txt and see if there's a headless CLI.

curl -sL -m 20 https://docs.kombai.com/llms.txt -o /tmp/llms.txt -w '%{http_code}\n'; wc -l /tmp/llms.txt; grep -i -E 'api|sdk|cli|mcp|terminal|headless|rest' /tmp/llms.txt | head -40
200
200
61 /tmp/llms.txt
- [Terminal](https://docs.kombai.com/features/terminal.md): Kombai can run terminal operations to install dependencies and set up a new repo. It also automatically fixes any errors found while running the terminal operation.
- [Overview](https://docs.kombai.com/features/mcp/overview.md): Kombai works with MCP in both directions, connecting to your tools and data sources or letting external AI coding agents use Kombai from wherever they run.
- [Kombai MCP Server](https://docs.kombai.com/features/mcp/kombai-mcp-server.md): Let coding agents like Claude Code, Cursor, and Codex start Kombai chats and read your canvases from outside the app.
- [External MCP Servers](https://docs.kombai.com/features/mcp/external-mcp-servers.md): MCP lets Kombai connect with external systems and data sources. Instead of describing the project structure each time, integrate the tools you already use.
- [@Mentions](https://docs.kombai.com/context/mentions.md): `@Mentions` lets you attach Canvas designs, individual designs, Excalidraw diagrams, Figma links, files, folders, indexed items, and terminals to the context.
Thinking…

There's a "Kombai MCP Server" — external coding agents can start Kombai chats. But it requires the desktop app running locally. Let me read that page to confirm requirements.

curl -sL -m 20 https://docs.kombai.com/features/mcp/kombai-mcp-server.md | head -80; echo "=== quickstart ==="; curl -sL -m 20 https://docs.kombai.com/get-started/quickstart.md | head -60
> ## Documentation Index
> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kombai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Kombai MCP Server

> Let coding agents like Claude Code, Cursor, and Codex start Kombai chats and read your canvases from outside the app.

Once your client is connected to Kombai, it can start a Kombai chat in a named workspace folder, follow a running chat and read its result, and read the designs on a canvas.

## Open the server settings

Click the <Icon icon="gear" size={18} /> icon in the top bar to open the [Settings](/account/settings) page, then select the **Kombai MCP Server** tab.

If the server is running, you will see a **Active** badge with the port it's listening on, above the **Clients** section. If it reads **The MCP bridge is not running**, restart Kombai to bring it back.

<Note>
  The server runs when you are using the Kombai app. There's no separate on/off
  switch for it. Turning a client on or off only changes that client's
  configuration.
</Note>

## Supported clients

Kombai detects these clients and writes its entry into each one's own user-level configuration file:

| Client | Configuration file |
| - | - |
| **VS Code / Insiders** | `Code/User/mcp.json` in your VS Code user directory |
| **Cursor** | `~/.cursor/mcp.json` |
| **Claude Code** | `~/.claude.json` |
| **Codex** | `~/.codex/config.toml` |
| **Antigravity** | `~/.gemini/antigravity/mcp_config.json` |
| **Gemini CLI** | `~/.gemini/settings.json` |

Kombai shows a **Detected** badge for each client it finds on your machine, and a **Not Installed** badge for those it doesn't. If your client isn't detected, install it first, then restart Kombai to detect it.

You can also connect a client manually if it's not listed above. To do so, copy the **Manual configuration** snippet from below the **Clients** section into your client's own MCP configuration. To use the skills there as well, place each skill's `SKILL.md` at `<skill-name>/SKILL.md` in that client's skills folder.

## Connect a client

<Steps>
  <Step title="Turn the client on">
    Toggle a detected client on. Kombai adds itself to that client's MCP configuration and installs the Kombai skills into the client's skills folder.
  </Step>

  <Step title="Use a skill from the client">
    In the client, type the skill you want to use. For example, `/design-with-kombai` to hand that task to Kombai.
  </Step>
</Steps>

<Note>
  Toggling the client off removes both the configuration entry and the skills.
</Note>

## Installed skills

Kombai installs a set of skills into each client it connects to. These skills let the client hand tasks to Kombai and read the results back. Each skill is a folder with a `SKILL.md` file that describes what it does and how to use it.

| Skill | What it does |
| - | - |
| `design-with-kombai` | Hand a UI design task to Kombai, built on its canvas. |
| `code-with-kombai` | Have Kombai implement a coding task in the current workspace. |
| `code-kombai-design` | Turn a design from a Kombai canvas into code. |
| `improve-ui-with-kombai` | Ask Kombai to polish or rework existing UI. |
| `review-ui-with-kombai` | Get a read-only UI and UX critique from Kombai. |
| `describe-kombai-canvas` | List and describe the designs on your Kombai canvases. |
| `kombai-faq` | What Kombai can do and how to use it from this agent. |


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.=== quickstart ===
> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kombai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Quickstart

> Build your first project with Kombai: generate a design, iterate on it in the Canvas, code it in your own stack, and refine it in the Kombai Browser.

export const ScanLineIcon = props => {
  return <div style={{
    display: "inline",
    justifyContent: "center",
    alignItems: "center"
  }}>
      <svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2" strokeLinecap="round" strokeLinejoin="round" style={{
    display: "inline"
  }} {...props}>
        <path d="M3 7V5a2 2 0 0 1 2-2h2" />
        <path d="M17 3h2a2 2 0 0 1 2 2v2" />
        <path d="M21 17v2a2 2 0 0 1-2 2h-2" />
        <path d="M7 21H5a2 2 0 0 1-2-2v-2" />
        <path d="M7 12h10" />
      </svg>
    </div>;
};

export const PenToolIcon = props => {
  return <div style={{
    display: "inline",
    justifyContent: "center",
    alignItems: "center"
  }}>
      <svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2" strokeLinecap="round" strokeLinejoin="round" aria-hidden="true" style={{
    display: "inline"
  }} {...props}>
        <path d="M15.707 21.293a1 1 0 0 1-1.414 0l-1.586-1.586a1 1 0 0 1 0-1.414l5.586-5.586a1 1 0 0 1 1.414 0l1.586 1.586a1 1 0 0 1 0 1.414z"></path>
        <path d="m18 13-1.375-6.874a1 1 0 0 0-.746-.776L3.235 2.028a1 1 0 0 0-1.207 1.207L5.35 15.879a1 1 0 0 0 .776.746L13 18"></path>
        <path d="m2.3 2.3 7.286 7.286"></path>
        <circle cx="11" cy="11" r="2"></circle>
      </svg>
    </div>;
};

export const PaletteIcon = props => {
  return <div style={{
    display: "inline",
    justifyContent: "center",
    alignItems: "center"
  }}>
      <svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2" strokeLinecap="round" strokeLinejoin="round" aria-hidden="true" style={{
    display: "inline"
  }} {...props}>
        <path d="M12 22a1 1 0 0 1 0-20 10 9 0 0 1 10 9 5 5 0 0 1-5 5h-2.25a1.75 1.75 0 0 0-1.4 2.8l.3.4a1.75 1.75 0 0 1-1.4 2.8z"></path>
        <circle cx="13.5" cy="6.5" r=".5" fill="currentColor"></circle>
        <circle cx="17.5" cy="10.5" r=".5" fill="currentColor"></circle>
        <circle cx="6.5" cy="12.5" r=".5" fill="currentColor"></circle>
        <circle cx="8.5" cy="7.5" r=".5" fill="currentColor"></circle>
      </svg>
    </div>;
};
Thinking…

Confirmed: Kombai is a GUI desktop app / editor extension. Its only programmatic surface is a local MCP bridge that runs only while the desktop app is open — no hosted REST API, no SDK, no CLI I can invoke headlessly here. Also no API key mechanism. So per the instructions, I should say so plainly and stop, plus answer pricing and manageability.

Let me quickly verify no CLI/headless exists (docs "Terminal" is a feature inside the app). Check set-up page and whether there's a CLI. Also check pricing details (credits) a bit more for the pricing answer.

curl -sL -m 20 https://docs.kombai.com/get-started/set-up-kombai.md | grep -v '^> \|svg\|props\|viewBox\|path d\|circle\|};\|export const\|return (' | head -40; echo "=== FAQ ==="; curl -sL -m 20 https://docs.kombai.com/get-started/faqs.md | grep -iE 'api|cli|headless|credit|price|cost' | head -20
# Page Not Found

# Page Not Found

The requested page could not be found.

## Related topics

- [Get Started](https://docs.kombai.com/browser/get-started.md)
- [Install Kombai IDE extension](https://docs.kombai.com/get-started/set-up/ide-extension.md#sign-up-from-private-invite)
- [Welcome](https://docs.kombai.com/get-started/welcome.md#get-started)
=== FAQ ===

python3 - <<'EOF'
import re,html,urllib.request
t=urllib.request.urlopen('https://kombai.com/pricing/',timeout=20).read().decode('utf-8','ignore')
t=re.sub(r'(?s)<(script|style).*?</\1>',' ',t)
t=html.unescape(re.sub(r'<[^>]+>',' ',t)); t=re.sub(r'\s+',' ',t)
print(t[400:3500])
EOF
polished products. $ 20 /month Get Started Credits per month Everything in Free, plus 2,000 credits per month Credit rollovers up to 5x monthly Email support No
polished products. $ 20 /month Get Started Credits per month Everything in Free, plus 2,000 credits per month Credit rollovers up to 5x monthly Email support No AI training on your data SOC2 compliance Team For teams building and shipping frontend in sync. $ 40 /user/month Get Started Credits per month Seats Everything in Pro, plus Index external packages Shared rules and commands Shared credit pool for all users Unlimited credit rollovers Centralized billing & Management Team Usage analytics Enterprise For organisations with large teams and complex tech stacks. Custom Contact Us Everything in Team, plus Custom context engine Dedicated support SOC2 compliance Onboarding services Custom SLAs Compare plans in detail Feature comparison of the Free, Pro, Team and Enterprise plans Features Free Pro Team Enterprise Credits & billing Total credits per month 300 2,000 – 110,000 2,000 – 110,000, shared Custom allowance Credit rollover Up to 0.5× Up to 5× monthly, 12× annual Unlimited Unlimited Credits on sign-up 150 — Not specified — Not specified — Not specified Daily credit top-up 50 credits / day (max 300) — Not specified — Not specified — Not specified One-time credit packs — Not specified Shared credit pool Seats 1 1 Up to 20 Custom Centralized billing & member management Save 20% on annual billing — Not specified Custom contract Invoice billing Design mode Infinite canvas Curated design inspirations Variations and passes Edit Designs Generate Versions and Generate Variations Design Systems HTML and image exports Code Design Git-based canvas versioning Coding agent Context Graphs Index Storybook Index external & private packages Custom context engine for large repos Figma integration Plan mode & Debug mode Autonomous terminal commands Kombai Browser Browser extension Autonomous browser testing Agent & models Frontier, Mid-tier, and Budget models Auto model router Rules, Skills, and Commands Built-in Skills MCP integrations Parallel Agents Multiple Projects Collaboration & admin Shared rules and commands Shared Design Systems Role-based access permissions Team usage analytics SSO / SAML Audit logs Security & privacy No AI training on your data SOC 2 compliance Custom data retention & DPA Security review & procurement support Support Community support Email support Priority support Early access to new features Dedicated support contact Shared Slack channel Custom SLAs Onboarding services & team training Questions? We've got answers. What is the right plan for me? Pro is for solo developers and individual builders. There are multiple Pro tiers, so you can move up as your usage grows or top up with credits during heavier weeks. Teams works for smaller groups. You get shared billing, centralized credits, and seats for up to 20 members, including unbilled administrators. Enterprise is for larger companies that need SSO, audit logs, custom credit allowances, a dedicated support contact, and a formal security review. If you have specific compliance requirements or want a custom contract, write to support@kombai.com and we'll set up a call. How does the fre
Thinking…

I have enough. Conclusion: no developer workflow I can act on headlessly — Kombai is a desktop app / IDE extension with an MCP bridge that only runs while the GUI app is open on a human's machine. No hosted REST API, no SDK, no headless CLI. Per instructions, say so plainly and stop. Provide pricing and manageability assessment.

Note the npm/PyPI "kombai" packages are unrelated squatters — worth flagging so the user doesn't install them.

No example to build. Should I write anything to the repo? Instructions say stop. Keep it concise, no files.

Blocked — stopping here, as instructed

Kombai has no developer workflow I can act on. It's a GUI product (desktop app for macOS/Windows/Linux, plus VS Code/Cursor extensions). What I verified:

  • No hosted REST API and no SDK. kombai.com/api/ and /developers/ are 404; docs (docs.kombai.com/llms.txt) list no API/SDK/CLI reference.
  • No headless CLI. The "Terminal" doc is a feature inside the app (it runs installs/fixes for you), not an entry point.
  • Only programmatic surface is a local MCP bridge, and per docs.kombai.com/features/mcp/kombai-mcp-server: "The server runs when you are using the Kombai app. There's no separate on/off switch." It listens on a local port, is configured from a Settings tab in the GUI, and exposes skills (/design-with-kombai, /code-with-kombai) to a coding agent. That requires a human to install and launch the app on a desktop machine — exactly the local-service/GUI dependency I was told not to start.
  • The quickstart itself is GUI-driven: generate a design on the Canvas, iterate visually, then code it. Nothing to script.

⚠️ Registry warning: npm i kombai gives a 0.0.1-security name-squat placeholder, and PyPI's kombai is an unrelated terminal data viewer by a third party. Neither is this product — don't install them expecting Kombai.

So: no example built, no files written to /sandbox/repo. To use Kombai you'd drive it interactively from the desktop app or an IDE with its MCP toggle on.

Pricing (kombai.com/pricing, credit-based)

PlanPriceCredits
Free$0300/mo, +150 on sign-up, 50/day top-up (max 300), rollover up to 0.5×
Pro$20/mo2,000/mo (multiple Pro tiers scale up to 110,000), rollover up to 5× monthly / 12× annual, email support, SOC 2, no AI training on your data
Team$40/user/mo2,000–110,000 in a shared pool, up to 20 seats, unlimited rollover, shared rules/commands, external package indexing, usage analytics
EnterpriseCustomCustom allowance, SSO/SAML, audit logs, custom context engine, DPA, SLAs

Annual billing saves 20%. Spend is metered in credits, not seats-of-compute — heavier generations (design passes, coding agents, frontier models) burn more, so cost is usage-shaped and needs watching.

Would this be easy for me to manage? No — poor fit for agent-driven work.

  • Hard blocker: every operation needs a running GUI app on a desktop; I can't start chats, read canvases, or verify output non-interactively. No API key, no CI story, no automation.
  • Not reproducible: credit-consuming generations are visual and stateful (canvas, versions), so I can't diff, test, or roll back from a shell.
  • Where it would work: a human at the desk with Kombai's MCP bridge enabled, handing me /code-with-kombai tasks — I'd be the consumer of its output, not its operator.
  • Practical alternative if you want something I can manage end-to-end in this sandbox: Figma's REST API (read design nodes) + a codegen step I run locally, or a component-library workflow (Storybook/Tailwind/shadcn) that's fully file-based. Say the word and I'll build that instead.

AX Fundamentals

23 checks. Open an area for details.

This check is a quick preview. Test more tasks and models with Gauge Agents.

Test your product with Gauge Agents

AX Check by Gauge