canicode

library0.12.3jsnpmunverified

CanICode v0.12.3 is a CLI and MCP server that lints Figma designs for AI code-generation readiness. It runs 16 deterministic rules (no AI cost) across 6 categories (Pixel Critical, Responsive, Code Quality, Token Management, Interaction, Semantic) and automatically fixes issues via Figma API, enabling a roundtrip workflow. Requires Node.js >=22. Differentiators: rule scores are ablation-validated with 94% pixel accuracy using 5× fewer tokens than raw JSON, and the calibration pipeline uses pixel-level comparison (visual-compare) against community fixtures. Supports both capture-only (gotchas) and write-to-Figma (roundtrip) modes. Released frequently (every few weeks) under active development.

devops testing web-framework

npm install canicode

github ↗package ↗

INSTALL

IMPORT

SIG · CANICODE

canicode

devopsjavascriptv0.12.3

harness data pending

Install & Compatibility

Where this runs

No compatibility data collected yet for this library.

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

default

✓ import canicode from 'canicode'

✗ const canicode = require('canicode')

ESM-only package; requires Node >=22. No default export in current API; prefer named CLI usage.

Analyzer

✓ import { Analyzer } from 'canicode'

✗ import Analyzer from 'canicode/Analyzer'

Named export for programmatic analysis. TypeScript types are bundled.

Roundtrip

✓ import { Roundtrip } from 'canicode'

✗ import { Roundtrip } from 'canicode/roundtrip'

Named export for the roundtrip (write-to-Figma) workflow. Introduced in v0.11.0.

Shows programmatic usage to analyze a Figma file and print grade, rule results, and gotchas.

#!/usr/bin/env node import { Analyzer } from 'canicode'; const figmaToken = process.env.FIGMA_TOKEN ?? ''; const fileKey = process.env.FIGMA_FILE_KEY ?? ''; const analyzer = new Analyzer({ token: figmaToken }); async function main() { const result = await analyzer.analyzeFile(fileKey); console.log('Overall grade:', result.grade); console.log('Rule results:', JSON.stringify(result.rules, null, 2)); console.log('Gotchas:', result.gotchas); } main().catch(console.error);

canicode --version

Debug

Known issues

breakingv0.12.0 removed the '--api' flag from 'analyze' subcommand. Use 'FIGMA_TOKEN' env var or config file instead.

fix

Set FIGMA_TOKEN environment variable or use 'canicode init --token' to configure.

affects: >=0.12.0

breakingv0.12.0 split 'canicode init' into interactive setup and separate 'token-refresh' command. The old non-interactive 'init --token' is removed.

fix

Run 'canicode init' interactively or use 'canicode token-refresh <token>' separately.

affects: >=0.12.0

deprecatedv0.11.0 deprecated direct export of 'RuleScore' type; use canonical types from '@canicode/types' if needed.

fix

Import from '@canicode/types' instead of 'canicode' for type-only imports.

affects: >=0.11.0

gotchaThe roundtrip writes to Figma may produce '📝 annotated the scene node' markers; this is correct behavior when writing to instance children. It does not mean failure.

fix

Review the annotation format; use 'allowDefinitionWrite' opt-in to propagate to component definitions.

affects: >=0.11.0

Errors

Common errors & fixes

Error: Cannot find module 'canicode'

Installed as global but used as local import without adding to package.json or running from wrong directory.

fix

Run 'npm install canicode' in your project and ensure node_modules contains it.

TypeError: analyzer.analyzeFile is not a function

Using outdated version <0.11.0 that did not export 'Analyzer' class.

fix

Update to v0.11.0+ with 'npm install canicode@latest'.

Figma API error: 403 Forbidden

Insufficient Figma token permissions (read-only token used for write operations).

fix

Generate a new personal access token with 'write' scope in Figma account settings.

Upgrade

Version history

0.12.3latest on npm

Audit

Dependencies

figma-apirequiredUsed to interact with Figma REST API for reading and writing design data

Agent activity

15 hits · last 30 days

node

Amazon

OpenAI (training)

Resources

githubhttps://github.com/let-sunny/canicode ↗

packagehttps://www.npmjs.com/package/canicode ↗