wrangler deploy failed: workers.api.error.script_too_large

wrangler deploy failed: workers.api.error.scripttoolarge

TL;DR

The worker bundle exceeds Cloudflare's size limit. Shrink it: check what's bloating the bundle (source maps, huge deps, inlined assets), externalize or trim, and redeploy. The limit is on the compressed bundle; measure that.

workers.api.error.script_too_large

Steps

1. Measure the bundle: check wrangler's deploy output for the reported size vs the limit. Expected: You know the overage.

2. Find the bloat: analyze the bundle (node_modules weight, inlined assets, duplicate deps). Expected: The biggest contributors are named.

3. Shrink it: remove unused deps, import subpaths instead of whole libraries, move large assets to R2. Expected: The bundle drops under the limit.

4. Redeploy and confirm the size in the output. Expected: The deploy succeeds.

5. Add a bundle-size check to CI so it cannot regress silently. Expected: Future bloat is caught pre-deploy.

When to use

• wrangler deploy fails with scripttoolarge
• the bundle grew recently
• large dependencies were added

When not to use

• the error is a different deploy failure (read it)
• the bundle is small (check for a misreported size)
• using Pages (different limits)

Tool compatibility

• wrangler
• Cloudflare Workers

Variant phrasings

• "workers.api.error.scripttoolarge"
• "wrangler deploy script too large"
• "worker bundle exceeds size limit"

Why it happens

Workers have a script size cap, and bundles grow silently as dependencies accumulate. A single heavy import can push a previously fine worker over. The error names the API; the fix is always making the bundle smaller.

Edge cases

• the limit applies to the compressed size; measure that, not the source
• source maps count if uploaded; keep them out of the deploy
• monorepo shared deps are the usual surprise contributor

Provenance

Resolved from the public thread: https://vectle.com/posts/pstlT4Rua8a8FB5-3JemIYxA