Configuration

edgefit works without a config file. Add one when you want to set targets once, ignore findings you have accepted, or change how strict a category is.

The config file

Create edgefit.config.ts in the project root. .mts, .js and .mjs work too.

import { defineConfig } from 'edgefit';

export default defineConfig({
  targets: ['workerd', 'bun'],
  entry: 'src/index.ts',
  ignore: [{ package: 'chokidar', reason: 'dev server only' }],
  levels: { unknown: 'off' },
});

defineConfig does nothing at runtime. It only gives you types and editor completion. The config is bundled with esbuild before it is loaded, so TypeScript and relative imports work on any supported Node version.

To use a config somewhere else, pass --config path/to/edgefit.config.ts.

Command-line flags always win over the config file.

Targets and entry

export default defineConfig({
  targets: ['workerd', 'deno'],
  entry: 'src/index.ts',
});

• targets defaults to ['workerd']. The valid names are workerd, bun, deno and deno-deploy.
• entry is relative to the project root. Without it, edgefit uses wrangler's main. One entry is scanned for every target, so a project with a wrangler config can check Bun and Deno without repeating it.

Ignoring findings

Some findings are fine: a package that is only loaded in development, or a code path you know never runs on the edge. Ignore them with a reason, so the next person knows why.

export default defineConfig({
  ignore: [
    // Every finding in one package
    { package: 'chokidar', reason: 'only imported by the dev server' },
    // One API everywhere
    { api: 'node:fs.watch', reason: 'guarded by a runtime check' },
    // Every member of a module, in one package
    { package: 'pg', api: 'node:net.*', reason: 'we use the Hyperdrive socket instead' },
    // Your own code
    { package: '.', api: 'node:child_process*' },
  ],
});

• package matches the package name. Use . for your own code.
• api matches the API as it appears in the report. A trailing * matches a prefix.
• A rule with both matches only when both match.

Ignored findings are counted in the report header and in the JSON output as ignored, so they do not disappear without a trace.

Levels

Each category has a default level. Set any of them to error, warning or off.

export default defineConfig({
  levels: {
    mocked: 'warning',
    mismatch: 'error',
    unknown: 'off',
  },
});

Category	Default
unsupported	error
mocked	error
mismatch	warning
web	warning
unknown	warning

Only error-level findings make edgefit check exit with 1.

Runtime settings

Cloudflare Workers

export default defineConfig({
  workerd: {
    compatibilityDate: '2026-04-01',
    compatibilityFlags: ['nodejs_compat'],
    wranglerConfig: 'wrangler.production.jsonc',
  },
});

By default these come from wrangler.jsonc, wrangler.json or wrangler.toml in the root, then from the data's own settings. Set wranglerConfig: false to ignore the wrangler config.

Bun

export default defineConfig({
  bun: { version: '1.4.2' },
});

By default the version comes from packageManager: "bun@x" in package.json, or from .bun-version. It is only used to warn you when your Bun is older than the data.

Deno

export default defineConfig({
  deno: { configFile: 'deno.jsonc' },
});

Used by both deno and deno-deploy. By default edgefit looks for deno.json, then deno.jsonc, in the root. Set configFile: false to skip the import map.

Export conditions

Add conditions that are checked before the target's own, for example when your bundler is configured with a custom one:

export default defineConfig({
  conditions: ['edge-light'],
});

See the config reference for every option in one place.