Lingua Detective vs CLD3: Chrome's language detector vs a crawled compliance report
Verdict: If you need a fast, compact, offline language detector — the kind that powers Chrome's own translate prompt — CLD3 is free and battle-tested; use it directly in your own code. If you need to know whether every page of a multilingual website is actually in the language it claims to be, you need crawling, JS rendering, and a verdict per page wrapped around that detector — which is what Lingua Detective adds.
| CLD3 | Lingua Detective | |
|---|---|---|
| What it is | Google's Compact Language Detector v3 — a small neural network for language ID, built into Chromium | Managed service that crawls a site and audits every page's actual language |
| Setup | Build the C++ library or install a binding (e.g. pycld3); expect some build tooling and protobuf-version friction | Give it a URL or a list of URLs — no compiling, no code |
| Pricing | Free, no account | Free trial (20 pages, no account/card); then $0.015→$0.010→$0.006/page, +$0.005/page if it crawls for you; or pay-per-call API from $0.002 |
| Privacy / offline | Fully offline — the same detector Chrome runs locally | Offline detection engine by default; uncertain pages can optionally go to AI arbitration |
| Output | A language code and reliability score for a string you supply | A per-page verdict — Conforming, Non-conforming, Uncertain, Error, or Undetermined — plus a site report |
| Best for | Developers who want Chrome-grade detection embedded in their own app | Teams that need an auditable view of a whole multilingual site, not a single string |
Pick CLD3 if…
- You want the same compact, neural detector that ships inside Chromium, embedded directly in your own application.
- You're comfortable building the C++ library or wrestling a Python binding's toolchain and protobuf-version requirements.
- You're passing it text you've already extracted, not asking it to find and render pages itself.
- You want zero licensing cost and are fine with a mature project that isn't under heavy active development.
- You need a lightweight, fast call inside a larger pipeline you already own and run.
Pick Lingua Detective if…
- You need a verdict for every page of a real site, including JavaScript-rendered ones, not a single string you've already pulled out.
- You'd rather skip compiling C++ or debugging a binding's build errors and just point at a URL instead.
- You want a report — compliant/non-compliant/uncertain per page — that a non-developer stakeholder can act on directly.
- Ambiguous pages should get a smarter second look instead of a raw confidence score you have to interpret yourself.
- You want scheduled, recurring audits without owning a crawler or the detector's build pipeline.
Pricing math
Auditing a 500-page site with URLs supplied: 200 pages at $0.015 ($3.00) plus 300 pages at $0.010 ($3.00) totals $6.00. Add Lingua Detective's crawl spider at $0.005/page (+$2.50) if you want it to discover the URLs itself, for $8.50. CLD3 itself is free, but compiling it, wiring up a crawler and renderer around it, and maintaining that stack over time is the cost Lingua Detective is replacing.
FAQ
Q: Is CLD3 hard to install?
It's a C++ library; Python bindings like pycld3 need a working compiler and matching protobuf version, which trips up some setups.
Q: Does Lingua Detective use CLD3 internally?
Lingua Detective runs its own offline detection engine, then adds crawling, JS rendering, and per-page verdicts that CLD3 alone doesn't provide.
Q: Is CLD3 still actively developed?
It's mature and widely deployed (it ships inside Chromium) but sees infrequent updates — typical for a stable, solved-enough component.