Dev Encyclopedia
ArticlesTools

Get notified when new content drops

No spam. Just new articles, tools, and updates straight to your inbox.

Dev Encyclopedia

A reference for builders

Content

  • Articles
  • Tools
  • Contact

Connect

  • support@devencyclopedia.com
  • RSS Feed

© 2026 Dev Encyclopedia

Privacy PolicyTermsDisclaimer
  1. Home
  2. /
  3. Tools
  4. /
  5. tsgoCalc
Free · Live

tsgoCalc: estimate your TypeScript 7 build time improvement

TypeScript 7 is 10x faster for VS Code (1.5 million lines). Your project is not VS Code. Enter your file count, lines of code, and CPU cores to see a speedup estimate, projected CI time savings, and whether it's worth trying the beta today.

New to TypeScript 7? Read the TypeScript 7 migration guide for install steps, breaking changes, and tsconfig updates.

  1. 1

    Enter your project size

    file count and lines of code range
  2. 2

    Set cores and CI frequency

    monorepo, CPU cores, runs per day
  3. 3

    Read your estimate

    build times, savings, and a recommendation

How tsgoCalc works

  1. 1

    Enter your approximate TypeScript file count (.ts and .tsx files, excluding node_modules), from 10 to 10,000+.

  2. 2

    Pick the band that covers your total lines of TypeScript code. This is the main driver of your baseline tsc build time.

  3. 3

    Tell it whether your project is a monorepo with independently type-checkable packages. tsgo's parallel checker benefits most from independent units of work.

  4. 4

    Pick the number of CPU cores on the machine where you'd run tsgo, typically your laptop for local development or your CI runner for pipeline times.

  5. 5

    Set how many times per day you run a full type check, counting both local runs and CI jobs.

  6. 6

    Read the results: estimated tsc and tsgo build time ranges, the speedup ratio, time saved per run, projected monthly CI hours saved, and (if you enter a CI cost) dollar savings, plus a recommendation.

What the output means

Estimated tsc (TS 6) / tsgo (TS 7) build time

The two ranges show a full, cold type-check of your project under the current compiler (tsc) and under TypeScript 7's Go-based compiler (tsgo). These are estimates of a clean tsc --noEmit / tsgo --noEmit run, not incremental rebuilds, which are faster on both compilers.

Estimated speedup

The ratio of your tsc time to your tsgo time. This number grows with project size and CPU cores: small projects on few cores might see 1.5-3x, while large monorepos on many cores can approach the 8-10x figures reported for Sentry and VS Code.

Time saved per run / monthly CI time saved

Time saved per run is the difference between your tsc and tsgo estimates. Monthly CI time saved multiplies that by your type-check frequency and 30 days, giving you a number you can use to justify the migration to a team or manager.

Estimated monthly CI cost saved

Only shown if you enter a cost per CI minute. Most CI providers publish a per-minute compute rate; multiplying that by your monthly time savings gives a rough dollar figure for the migration's ROI.

Recommendation badge

One of four outcomes based on your inputs:

  • Try the beta today: your build time is long enough that the speedup is worth capturing now.
  • Wait for stable: the speedup is real but modest, so there's little urgency before the stable release.
  • Plan a Q3 migration: a large monorepo where the gain is significant but a coordinated, package-by-package rollout makes sense.
  • Low urgency, small project: your project already type-checks quickly; migrate whenever convenient.

Installing tsgo to check your real numbers

Once you have an estimate, the only way to get an exact number is to run tsgo on your own project. The native preview package works alongside your existing tsc setup, no migration required to try it.

# Install the native preview (npm)
npm install --save-dev @typescript/native-preview@beta

# Or with pnpm
pnpm add --save-dev @typescript/native-preview@beta

# Time a full type check with tsc (TS 6)
time npx tsc --noEmit

# Time the same check with tsgo (TS 7)
time npx tsgo --noEmit

For tsconfig changes needed before upgrading, the TSConfigBuilder tool generates a tsconfig.json compatible with TypeScript 7 defaults. For the full list of breaking changes and a step-by-step plan, see the TypeScript 7 migration guide.

When to use tsgoCalc

ScenarioWhat to enter
"My CI type check takes 4 minutes, what would TS7 save me?"Your real file count and LOC band, your CI runner's core count, and your daily run count, plus your CI cost per minute for a dollar figure
Pitching the migration to your teamYour project's real numbers, then use the monthly CI hours/cost saved as the headline figure in your proposal
Deciding whether to install the beta todayYour current numbers, read the recommendation badge (try now vs wait for stable)
Sizing a large monorepo migrationMonorepo set to yes, your highest LOC band, and your CI runner's core count to see if the gain justifies a Q3 migration plan
Comparing a small app vs a large oneRun the calculator twice with different LOC bands to see how the speedup scales with project size
Checking if it's worth upgrading a small side projectA small LOC band and low CI run count, expect a "Low urgency" recommendation

Frequently Asked Questions

What does tsgoCalc estimate?

tsgoCalc estimates how much faster your TypeScript project's type-checking would be under TypeScript 7's Go-based compiler (tsgo) compared to the current TypeScript 6 compiler (tsc).

Enter your approximate file count, total lines of code, whether your project is a monorepo with independent packages, how many CPU cores your dev machine or CI runner has, and how often you run a type check. tsgoCalc returns an estimated tsc build time, an estimated tsgo build time, a speedup ratio, projected monthly CI time savings, and a recommendation for whether to try the beta now, wait for the stable release, or plan a phased migration.

How does tsgoCalc compare to just running tsgo myself?
tsgoCalcRunning tsgo --noEmit
Setup timeNone, instant estimate in your browserInstall @typescript/native-preview, then run it
AccuracyApproximate, based on a regression modelExact, measured on your real codebase
CI savings projectionBuilt in, based on your CI run frequencyYou compute it yourself from before/after times
Migration recommendationIncluded (try now / wait / plan)Not provided
Best forDeciding whether it's worth installing the betaConfirming the real number once you've decided

tsgoCalc is meant as a first pass: a way to decide whether installing a beta compiler is worth 10 minutes of your time today. Once you've installed it, the real number from your own CI logs will always be more accurate than any estimate.

Is my project information sent to DevEncyclopedia's servers?

No. tsgoCalc runs entirely as client-side arithmetic in your browser. There are no network requests, no uploads, and nothing is stored or transmitted. You're only entering rough numbers (file count, LOC range, core count), never actual source code.

How do I find my project's actual file count and line count?

Run these from your project root. They count only TypeScript and TSX files, excluding node_modules and build output:

bash
# File count
find . -name "*.ts" -o -name "*.tsx" | grep -v node_modules | grep -v dist | wc -l

# Total lines of code
find . -name "*.ts" -o -name "*.tsx" | grep -v node_modules | grep -v dist | xargs wc -l | tail -1

If you have cloc installed, cloc --include-lang=TypeScript . gives a cleaner breakdown including comment and blank line counts.

Why does my project show a smaller speedup than VS Code's 10x?

The 10x figure comes from VS Code's 1.5 million line codebase running on a machine with many CPU cores. tsgo's biggest wins come from parallel type-checking across independent modules, so the speedup scales with project size and core count.

A smaller project (say, 50,000 lines on a 4-core laptop) has less independent work to parallelize and a smaller baseline build time to begin with, so a 2-4x speedup is typical and still meaningful for editor responsiveness, even if it's far from 10x.

When will TypeScript 7 be stable, and should I wait?

TypeScript 7.0 stable is expected in late June or early July 2026. The Beta (available as @typescript/native-preview) is described by Microsoft as stable enough for daily use and CI pipelines.

If tsgoCalc's recommendation says "Try the beta today," your project is large enough that the speedup is worth capturing now. If it says "Wait for stable," the gain is real but modest enough that there's little downside to waiting a few weeks. See the full migration guide for the breaking changes to check before upgrading either way.

Related reading

Guide

How to Migrate to TypeScript 7 (tsgo): Breaking Changes and Build Times

Install steps, the four real breaking changes, tsconfig updates, and a decision framework for whether to upgrade now.

Article

TypeScript 7 (Project Corsa): What Next.js Devs Need to Know

The original announcement explainer: why Microsoft rewrote the compiler in Go and what changed.

10,000+
Total lines of TypeScript (approx.)
Monorepo with independent packages?
CPU cores available (dev machine or CI runner)
10

Estimated tsc (TS 6) build time

9.1s - 12.8s

Estimated tsgo (TS 7) build time

1.8s - 2.5s

Estimated speedup

5.1x faster

Time saved per run

8.6s

Monthly CI time saved

43 min

Recommendation: Wait for stable

The speedup is real but moderate at this size. It's reasonable to wait for the TypeScript 7 stable release (expected late June or early July 2026) rather than running beta tooling in CI today.

How we estimate this

The tsc baseline is a curve fit to four published benchmarks: a ~7,500-line app (3.2s), a ~100K-line backend (12s), the ~600K-line Sentry monorepo (133s), and the 1.5M-line VS Code codebase (78s). Build time grows slightly slower than linearly with lines of code, reflecting incremental gains from module caching at larger scales.

The raw tsgo speedupfollows a log-linear curve fit to the same four points (each point lands within about 6% of the published number). Larger codebases see a bigger speedup because the Go compiler's parallel, goroutine-based type checker has more independent work to spread across cores.

That raw speedup is then adjusted for your inputs: CPU cores apply a diminishing-returns multiplier (2 cores ~1.8x, 4 cores ~2.8x, 8 cores ~4.1x, 16+ cores ~5.5x), normalized against an 8-core baseline. A monorepo with independent packages captures the full multiplier; a single package captures about half of it, since there are fewer fully-independent units to check in parallel.

Finally, file countis compared against a typical ~250 lines-per-file baseline for your selected LOC range. More, smaller files give tsgo's checker finer-grained parallelism (up to +30% speedup); fewer, larger files reduce it (down to -30%).

This produces an estimate, not a measurement. The only way to know your project's real number is to install @typescript/native-preview and run npx tsgo --noEmit yourself.

All calculations run locally in your browser. Nothing is sent to DevEncyclopedia servers.