Starter
For small localization teams of 2–4 people.
$399/ month
Billed yearly: $4,788 / year
- 1 million words included
- 3 user licenses
- Shared cloud hosting included
- AI translation billed by usage
Pricing
Pick the plan that fits your team. Every plan includes the full CENTROID platform and secure cloud hosting. AI translation is billed only for the words you actually run.
A yearly planUsers, included word volume, the platform and hosting.
AI by usage$600 per 100,000 words, or $5,000 per 1 million words.
Only the extras you needDedicated hosting, faster support, Screenshot Review, connectors.
For small localization teams of 2–4 people.
Billed yearly: $4,788 / year
For small and mid-sized, growing companies.
Billed yearly: $11,988 / year
For large organizations with high volume.
Billed yearly: from $29,988 / year
Not sure yet? Start with a 15-day pilot for $1,000.A guided pilot (about 2 weeks) on your real content, before you choose a plan.
Pick a plan and how many words you expect the AI to translate. The total updates as you go.
AI usage is charged at $600 per 100,000 words, or $5,000 per full million, whichever is cheaper.
An estimate in US dollars, before GST/VAT. Screenshot Review and connectors are priced separately below.
| Feature | Starter | Growth | Enterprise |
|---|
Extras you can add to any plan.
Check translations in context, on real screenshots of your product, before you ship.
AI reviews each screenshot and automatically retranslates anything that doesn't fit or reads wrong.
AI finds the issues and a language expert checks and fixes them. Includes right-to-left (RTL) languages.
Connect CENTROID to the tools where your content lives. Prices are per connector instance (one connected site or repository).
| Connector | Price | 2–5 instances | More than 5 |
|---|---|---|---|
| WordPress | $2,000 per instance | 50% off each $1,000 per instance | 70% off each $600 per instance |
| Git / GitHub | $4,000 per instance | 50% off each $2,000 per instance | 70% off each $1,200 per instance |
| Adobe Experience Manager (AEM) | $10,000 | Talk to sales for multiple instances | |
| Figma | $1,000 | Includes any number of instances | |
Coming soon: Sitecore, Marketo, Drupal, Contentful, HubSpot and SharePoint. Ask our sales team for early access.
Included in every plan
Secure, managed hosting on CENTROID's shared cloud. Nothing to set up.
or $10,000 / year Save $2,000
or $8,000 / year Save $400
Complex enterprise support needs?Our sales team will build a support plan with you.
FAQ
CENTROID goes beyond traditional TMS functionality by integrating AI-powered translation, automated quality checks, and MQM quality scoring. It supports fully automated, human-led, and hybrid workflows. With GitHub, REST API, CLI, and WordPress integrations, plus a browser-based CAT editor with live previews and real-time collaboration, CENTROID streamlines the entire localization process.
CENTROID uses project context, approved glossaries, translation memory, and surrounding sentences to improve translation accuracy. Automated checks identify missing text, incorrect numbers, script errors, formatting issues, and glossary mismatches. Failed checks trigger an AI retry or human review. MQM scoring and visual layout checks help teams identify errors, while human-approved translations take priority in translation memory.
Yes. CENTROID supports fully automated, human-led, and hybrid workflows, including AI translation, automated quality checks, human review, and proofreading. Project managers assign linguists and reviewers by document and language. Automatic approval can be based on minimum quality scores, while failed checks remain drafts for human review. GitHub translations are published only after reviewer approval.
Yes. CENTROID supports importing translation memories in TMX, CSV/TSV, JSON, and XLIFF/SDLXLIFF formats, and glossaries in TBX, CSV, TSV, or JSON. Existing translations and terminology can be reused from day one. Data remains portable through export options, while source content can be uploaded directly or synced through GitHub. A pilot helps validate migration before full rollout.
CENTROID provides separate client workspaces, role-based access controls, and restricted document and language permissions. Translation memories remain isolated between clients. Administrators can control AI access, allocate word credits, and monitor usage. Security measures include hashed passwords, signed session tokens, revocable API keys, and encrypted GitHub tokens. Templates standardize terminology, tone, AI instructions, and workflows across teams.
Yes. CENTROID analyzes translation memory matches to identify reusable content and reduce unnecessary translation costs. AI generates checked first drafts, allowing linguists to focus on review and editing. Teams can configure workflows to match their quality requirements. GitHub and API integrations also automate the localization of new or updated software and website content, reducing manual handoffs and delivery time.
You can evaluate CENTROID through a pilot using representative content, existing translation memories, glossaries, and target languages. Test translation quality, automated checks, workflow flexibility, integrations, and reviewer collaboration against your requirements. Compare results with your current TMS for accuracy, turnaround time, and cost. A pilot helps identify configuration needs and determine whether CENTROID fits your localization operations.
CENTROID supports more than 110 languages and regional variants, including major European, Asian, African, and South Asian languages. Supported locales include US and UK English, Canadian French, Latin American Spanish, Brazilian and European Portuguese, and Simplified and Traditional Chinese. It also supports right-to-left languages such as Arabic, Hebrew, Persian, and Urdu, alongside extensive Indian language coverage.
CENTROID implementation timelines depend on your content volume, integrations, workflows, and migration requirements. For a typical TMS migration, a one-week migration plan can serve as an initial planning benchmark, subject to the scope and complexity of the project. Larger enterprise deployments or complex integrations may require additional time.
Yes. CENTROID supports multiple workspaces, each with separate users, projects, translation memories, and settings. Within each workspace, teams can configure templates, languages, glossaries, tone, AI instructions, and workflows to suit their needs. Projects can target multiple languages, while GitHub and API integrations support continuous localization. Infrastructure capacity can be planned and scaled as enterprise usage grows.