Should You Leave Depcheck? 26 Files Knip Caught

On a 95-file JavaScript fixture, Knip 6.39.0 took a median 575 ms and depcheck 1.4.7 took 581 ms. Wall time was a wash. Knip also reported 26 unused files and 56 unused exports that depcheck does not look for, plus the same 8 unused production dependencies.

Short answer: leave depcheck for Knip when unused files and unused exports matter as much as unused packages. Keep depcheck when you only want a small package.json unused-deps check and nothing else. No affiliate links in this post.

What I tested

• Machine: 8 vCPU Intel Xeon / 15 GB RAM Linux cloud box, Node 20.19.2, tested 5 Oct 2026 (about 23:05 to 23:15 IST). Shared box; tools interleaved each round.
• Fixture: 95 files under src/ and scripts/ (libs, pages, dead orphans, a CLI entry). Production dependencies included used packages (lodash-es, chalk, ms) and eight deliberately unused ones (moment, left-pad, is-odd, bluebird, request, jquery, axios, and one more unused util package).
• Tools: knip 6.39.0 with knip.json entry points src/index.js and scripts/cli.js; depcheck 1.4.7 with defaults. Timed via direct node on each package bin.
• Timing: one warmup each, then five interleaved warm runs. Median wall-clock of the full CLI call.
• Findings: Knip unused files 26, unused dependencies 8, unused exports 56 (exit 1). Depcheck unused dependencies 8 (exit 255). Both tools agreed on the unused production dependency set.
• One surprise: speed was effectively tied, so the leave-depcheck case is coverage, not milliseconds.
• Not tested: TypeScript project references, monorepo workspaces, Next.js / Nuxt plugin presets, unused types, unused class members, CI baseline ignore files, Windows/macOS, false-positive rates on a real production repo.

Does Knip find more than depcheck?

Tool	Median	Runs (ms)	Unused deps	Unused files	Unused exports
Knip 6.39.0	575 ms	570 / 562 / 586 / 582 / 575	8	26	56
depcheck 1.4.7	581 ms	712 / 683 / 563 / 581 / 581	8	n/a	n/a

On this fixture both CLIs finished in about 0.58 s. Depcheck answered the package.json question. Knip answered that plus unused files and unused exports. If your backlog is dead modules and dead exports, depcheck alone will not surface them.

Dead-code cleanup pairs with lint and format work. The Biome vs ESLint + Prettier post and the oxlint vs ESLint post are the related quality-gate decisions on this site; this post is the unused-code inventory step. Package install speed is a separate question covered in the pnpm vs npm vs Bun post.

What do you lose leaving depcheck?

depcheck is smaller and single-purpose. Teams that only gate on unused dependencies, and that already ignore unused files elsewhere, may not want Knip's broader report noise. Knip also needs entry and project globs; a wrong entry list can mark live files as unused.

Also expect to maintain an ignore list. Orphan scripts, intentional barrels, and generated clients often need explicit allow rules before Knip is green in CI. Re-run both tools on your own tree before deleting packages or files.

Should you leave depcheck for Knip?

Situation	My pick
You care about unused files and unused exports, not only packages	Switch to Knip; keep an ignore list
You only want unused dependencies in package.json	Keep depcheck
Monorepo with clear entry points already listed for Knip	Knip is the better fit to evaluate
CI already fails on any unused-export noise you cannot afford	Stay on depcheck, or adopt Knip with a tight ignore
First PR would delete 26 orphan files without review	Inventory with Knip, then delete in small batches

Bottom line: on 95 files, Knip and depcheck both took about 0.58 s, and Knip also found 26 unused files and 56 unused exports. Who should not switch yet: teams that only want a package.json unused-deps gate. Everyone else cleaning dead modules should measure their own tree the same way.

How this was made: I built the fixture, ran interleaved Knip and depcheck on the ShopperCove test box, and kept the timing JSON; the write-up was drafted with AI help and checked against that output.

Sources

• 01knip.devhttps://knip.dev/
• 02github.com/webpro-nl/knip
• 03github.com/depcheck/depcheck
• 04npmjs.com/package/knip
• 05npmjs.com/package/depcheck

Related

• 01shoppercove.com/blog/eslint-10-12-release-october-2026
• 02shoppercove.com/blog/dprint-vs-prettier
• 03shoppercove.com/blog/typescript-7-should-you-upgrade
• 04shoppercove.com/blog/jest-vs-vitest
• 05shoppercove.com/blog/turborepo-2-11-7-sdkroot-oidc-cache-october-2026