Crowdin pricing breakdown: what you actually pay in 2026

Crowdin's plans run from free to $450/month, but the number that decides your bill is hosted words — and it multiplies by every target language. Here's the math, the gotchas, and who each plan fits.

8 min read← All articles

Crowdin's pricing page looks simple: a free plan, three self-serve tiers, and custom plans above that. The complexity is in the meter. Crowdin bills on hosted words, and hosted words multiply by every target language you add — which means the plan you need is a function of your content size times your language count, not either number alone.

Here's what each plan includes as of September 2026, where the bill grows faster than teams expect, and who each plan fits.

The plans, as of September 2026

Prices below are from crowdin.com/pricing as of September 2026. Crowdin offers monthly billing or annual billing with 2 months free — roughly 17% off, so the effective annual rates land near $49, $149, and $375 per month for the three self-serve paid tiers.

Prices checked September 22, 2026.

PlanMonthly priceWhat it adds
Free$01 private project and a starter hosted-word allowance
Pro$59/moMore private projects and integrations
Team$179/moLarger hosted-word allowance, more projects, manager seats
Team+$450/moLarger allowances again, plus more advanced workflows
Business / EnterpriseCustomCustom allowances and support; Crowdin Enterprise (cloud-hosted) has a 30-day trial

Each plan includes a set hosted-word allowance, and Crowdin's price calculator lets you add hosted words and managers on top. Allowances are set in that calculator, so check it for the numbers that apply to you.

A few genuinely good things about this structure: translators and proofreaders are unlimited on every paid plan (you pay for content and manager seats, not for the people doing translation work), there's a free plan plus a free license for open-source projects, a 14-day trial of the Team plan, and a 30-day money-back window. You can also add hosted words without jumping a full tier.

Annual vs monthly: the actual math

The annual discount is real but it's a prepay, so here's the effective math on the self-serve tiers:

PlanMonthly billingAnnual billing (prepaid)Effective monthly
Pro$708/year$590/year~$49/mo
Team$2,148/year$1,790/year~$149/mo
Team+$5,400/year$4,500/year~$375/mo

If you're confident in Crowdin after the 14-day Team trial, annual is the obvious call. If you're still comparing tools (Lokalise's restructured plans are the usual alternative — we've broken down Lokalise's 2026 pricing separately), stay monthly until the meter math below is settled, because the hosted-words model is where Crowdin budgets tend to surprise people.

The hosted-words math is the whole ballgame

Crowdin's own definition: hosted words are the number of words to be translated multiplied by the number of the project's target languages. The source language doesn't count.

Their example: upload a file with 500 words to a project with 10 target languages and you've used 5,000 hosted words.

Scale that to a real product and the multiplication does the work:

Your app's source contentTarget languagesHosted words
3,000 words (small SaaS)515,000
6,000 words (typical SaaS)1060,000
6,000 words1590,000
20,000 words (larger product)15300,000

Compare those totals with the allowances in Crowdin's calculator. The pattern to notice: adding a language counts your entire existing content again. Going from 10 languages to 15 on a 6,000-word app doesn't add 5 languages' worth of new strings — it adds 30,000 hosted words, which can mean a bigger allowance or a higher tier. Language expansion is where hosted-word bills grow, so model it before you commit, while switching is still cheap.

The other gotchas

Overage is a countdown, not a soft limit. Exceed your hosted-word quota and you get a 10-day grace period to upgrade. That's humane compared to a hard stop, but it means a big content push can force a tier decision on 10 days' notice.

Integrations are tier-gated. How many integrations you can connect (GitHub, Figma, and so on) depends on your plan, so check the count before you pick a tier. Some marketplace apps can be paid separately on top.

Manager seats are the hidden seat meter. Translators are unlimited on paid plans, but managers — the people who configure projects, workflows, and integrations — are counted, and you can add more at extra cost. On developer-led teams where every engineer is effectively a manager, this is the constraint you'll hit, not translator seats.

Annual pricing is a prepay. The "2 months free" discount requires paying the year upfront. Nothing unusual, but the advertised effective rates assume it.

And the cost that isn't on the pricing page at all: Crowdin manages strings that already live in translation files. If your codebase still has hardcoded strings, the extraction work — often days to weeks of developer time — comes first.

A worked example: year one for a typical SaaS team

Put the pieces together for a concrete team: a 6,000-word SaaS app (roughly 1,000-1,200 UI strings), 10 target languages, 4 developers, no dedicated localization staff, strings currently hardcoded.

Line itemCost
Extraction: finding and wrapping ~1,000 strings by hand$1,200-3,000 in developer time (one-time)
Crowdin Pro or Team, depending on the allowance 60,000 hosted words needs$708-2,148/year on monthly billing, or ~$590-1,790/year paid annually
Language expansion later (15 languages → 90,000 hosted words)A larger allowance, an add-on, or a higher tier
Year 1 total~$1,800-5,150

Two things stand out. First, the biggest single line may not be Crowdin — it's the extraction work that comes before it. Second, the hosted-words math means content growth or a new language can prompt a tier conversation. Neither is a scandal; both should be in the budget before you sign.

Who Crowdin is genuinely right for

An honest read, because Crowdin is widely used for good reasons:

  • Teams working with external translation agencies. Crowdin has a translator marketplace and vendor management built in. If you hire professional translators, this is the strongest reason to pay for Crowdin.
  • Enterprises with a localization manager. The workflow engine, reviewer roles, and multi-step workflows are built for someone whose job is running localization. If that person exists at your company, Crowdin is built around their day.
  • Mobile teams that need over-the-air delivery. Crowdin can push translation updates to shipped mobile apps without an app-store release. Polyglot doesn't do OTA; if you need it, that's a real differentiator.
  • Teams with unusual toolchains. 700+ integrations, and the ecosystem depth of a platform that's been around since 2009.

If two or more of those describe you, Crowdin's Team tier is fairly priced for what it does. See our full Polyglot vs Crowdin comparison for the feature-by-feature view.

How Polyglot's pricing compares

Polyglot is a different category of tool — a CLI that scans your source for hardcoded-string candidates, rewrites them into translation calls from a plan you review, translates, and previews locally, with generated translations reviewed in the dashboard. The pricing model is different in two structural ways.

First, the local workflow is free. Scanning (scan), planning and wrapping (start, wrap), and local preview (preview) are free and need no account.

Second, the meter is source strings, not words stored times languages. Translation is metered in source strings — a lifetime allowance that never resets — plus monthly character volume on Team and Scale. Adding a language to already-translated strings doesn't re-consume your string quota, because each distinct source string counts once (your plan's language limit still applies). Compare that to hosted words, where adding a language multiplies your usage.

Polyglot ProPolyglot TeamPolyglot Scale
Price$49/mo ($39 annual)$149/mo ($119 annual)$499/mo ($449 annual)
Strings (lifetime)10,00025,000Unlimited
Languages515Unlimited
CharactersUnlimited3M/mo25M/mo
Seats1515

Full details are on the pricing section of the homepage, and there's a free tier with 500 lifetime strings to migrate a real project before paying anything.

On quality — the thing a cheaper meter is worthless without — Polyglot published a re-judgeable benchmark of its June 2026 pipeline, AI-judged: 4.96/5 across 49 languages, 100% placeholder preservation, and 94.5% of translations given the top score (5/5), with the dataset and judging harness open source so you can re-score the June outputs yourself. Production has since changed models, so treat those as June's results.

The difference, stated plainly: Crowdin starts from strings already in translation files; Polyglot starts from the hardcoded strings in your code. If your team has a localization manager and agency translators, buy Crowdin. If your team is developers with a codebase full of hardcoded strings, run this first and see what you're actually dealing with:

curl -fsSL https://getpolyglot.ai/install.sh | bash
polyglot scan

The scan is free and doesn't require an account — which makes it a cheap way to size the work before you price out either tool.

Start in your terminal

Stop hunting for untranslated strings.

Install the CLI, run a scan, and see exactly what you're missing. Free, no account required.

$curl -fsSL https://getpolyglot.ai/install.sh | bash
Crowdin pricing breakdown: what you actually pay in 2026 - Polyglot Blog | Polyglot