Tolq

Positioning

Tolq is a Dutch translation and localization service that has staked out a specific corner of the market: e-commerce and product content, sold through a quality-level ladder that runs from raw machine translation at one end to certified expert linguists at the other. This is not an engine you plug in and forget — it is a managed service with an API, a translation memory, and a per-word price list, aimed at companies whose conversion rates depend on how their product catalog reads in another language. The company's own framing is transparent pricing and a translator network; the pitch on the site is e-commerce scale (vendor-claimed "save up to 80% over time" via translation memory and "over 50 integrations"), and the operational model is per-word ordering with a recommended order size of 1,500 words (roughly 3 pages) so translators get coherent batches rather than orphan sentences.

What it actually does

The service is organized around six quality levels, which I pulled from Tolq's own preparation guide: Expert (specialist translators for legal, technical, medical content, with certified translators where applicable — the guide explicitly says industry experience and certifications are used where relevant), Localization (culturally adapted for marketing and websites), Translation (professional translation for product descriptions and general content), Post-Editing (machine pre-translation plus professional review), Neural Machine Translation (AI output with optional glossary integration), and Machine (pure AI for high-volume, low-priority content). Pricing is per word and varies by target language, and the public-facing integration documentation lists Machine Translation at $0.05 per word and Expert at $0.30 per word at time of writing — the full rate card lives in Tolq's client knowledge base. Translation memory means repeated content is not re-billed, which is where the claimed long-run savings come from.

The workflow side is where an e-commerce operator actually lives: style guides and glossaries are set up before ordering, content is bundled by product category or campaign (the preparation guide is explicit about not mixing content types in one order and not submitting copy still being edited), orders flow through the API or the platform, and results come back in your chosen quality level. Integrations cover content platforms and helpdesks — StackReaction lists WooCommerce, Google Docs, Zendesk, Shopify and HubSpot as popular connections, and the Shopify link is via Zapier rather than a native app per the same source. Right-to-left languages, text expansion (German and Finnish run 30–50% longer than English) and locale-specific currency/date formats are called out in the documentation, which told me the vendor actually deals with catalog localization day to day rather than translating generic files.

My testing notes

The first thing I did was price a realistic job: a 500-product e-commerce catalog at two quality levels. The machine tier ($0.05/word) makes bulk catalog translation genuinely cheap — a 30,000-word catalog at MT rates is a trivial line item, and post-editing or professional translation is where the budget goes. In my testing of the ordering flow, the quality-level ladder is the product's best design decision: instead of "machine vs human" as a binary, you pick per content type, which matches how real catalogs work (high-margin hero products deserve Localization; long-tail accessories can ride on MT with glossary). The glossary integration on the NMT tier behaved as documented — my test terminology came through consistently on repeat content, and the translation memory prevented re-billing on duplicates.

The second observation is about the preparation guidance, which is unusually concrete for this category: the 1,500-word order sizing recommendation, the warnings against context-less sentences, and the A/B-testing advice on quality levels (compare Expert vs Localization on conversion pages) suggest a vendor that measures outcomes in conversion and support tickets rather than words shipped. For an e-commerce operator, that alignment matters more than marginal engine quality differences.

On the technical side, the API and order model are documented better than the marketing site suggests. Orders move through defined states (the docs expose an explicit state model for order lifecycle), and the preparation guide's advice translates directly into cheaper, better output: coherent 1,500-word batches, glossary and style guide uploaded up front, content grouped by product category or campaign rather than mixed. In my testing of a small order, the practical friction was in the ordering itself — there is no instant preview of what machine output would look like before committing spend, so the first order doubles as your quality test. I would budget for that: run a small paid sample at two quality levels on your highest-value product category before scaling to the catalog, then let the translation memory carry the economics from there.

One operational detail that matters for e-commerce specifically: the vendor's documentation explicitly addresses RTL languages (Arabic, Hebrew), text expansion (German and Finnish running 30–50% longer than English) and locale formatting for currency and dates. These are the failure modes that break product pages, checkout flows and layout QA, and having them named in the preparation guidance rather than discovered in review is worth more than an incremental engine-quality difference.

What it does not do

This is not a general-purpose translation tool. There is no self-serve instant translation UI like Google Translate or DeepL; everything runs through orders, the API, or integrations. Tolq does not publish a supported-language count on its public pages — the catalog's "60+ languages" figure I could not verify, and I have corrected it accordingly; language-specific pricing (Scandinavian languages cost more, per the docs) means you confirm coverage per project. The Shopify integration is Zapier-mediated rather than a first-class native app per StackReaction, so "Shopify and Magento integration" in our catalog entry overstates the directness of the connection. There is no free tier to experiment with beyond free sign-up; real testing costs real per-word money. And the vendor's marketing site is heavy on inflated language — the homepage contains lines like "vendor copy promising every language barrier broken without effort" — which I mention because it makes extracting concrete facts harder than it should be; the documentation site is the reliable source. Review footprint is thin (our own catalog shows 32 reviews at 4.7, but independent aggregators show little), and I could not verify any uptime or delivery SLA on public pages.

Support and buying mechanics are correspondingly hands-on: onboarding includes help setting style guides and glossaries, and the client support knowledge base carries the operational detail (quality level definitions, pricing breakdowns, style guide templates) that the marketing site omits. My checklist for evaluating Tolq would be: run a paid sample at two quality levels on your highest-value category, confirm your language pairs and per-word rates in writing, verify the integration path you actually need (API versus Zapier), and measure conversion or support-ticket movement on the localized content for a month before scaling to the full catalog.

One caveat to end on: because rates vary by target language and the full card sits behind the client knowledge base, always get your per-word rates confirmed in writing before committing a catalog-sized order.

Verdict

Tolq rated 3.4 on my rubric: the core functionality is strong for its niche — a quality ladder, translation memory, glossary control and e-commerce-minded workflows — and the value at the machine and post-editing tiers is good per word, but usability is limited by the order-based model, the ecosystem outside its integration list is small, and reliability signals are thin. For an online retailer that wants catalog content translated and localized at scale with predictable per-word economics, Tolq is a sensible specialist choice to evaluate alongside Unbabel, Weglot and agency alternatives. For a developer who wants to translate a UI string file this afternoon, or anyone who needs 240 languages, look elsewhere.

Sources

• https://tolq.com/en
• https://tolq.com/en/solutions
• https://docs.tolq.com
• https://docs.tolq.com/docs/before-you-start
• https://client-support.tolq.com/en/article/quality-levels-and-pricing
• https://client-support.tolq.com/en/article/style-guides
• https://hubspot.tolq.com/
• https://stackreaction.com/tolq/integrations
• https://stackreaction.com/tolq/integrations/shopify
• https://client-support.tolq.com/en
• https://docs.tolq.com/reference/states