We translate websites and stores: a target-language translator edits a draft from language models or DeepL. We adapt content to the market, researching queries and preparing meta tags, image descriptions and hreflang.
Four things always included in your Balance IT translation package.
Page / text brief + target language → quote within 24h.
Language models or DeepL prepare a draft, then a native translator edits it. More than machine translation alone.
Meta title + description for each language + keyword research for each market + hreflang + canonical.
Your TM (Translation Memory) — consistency over time. Switching providers = TMX export.
Machine-only translation is literal. Human-only translation is expensive + slow. AI draft + human edit is optimal: fast + affordable + local.
AI draft (language models / DeepL) — 80% speed
Native translator editing — tone + local idiom + SEO
Translation Memory (Phrase / Lokalise) — consistency over time
AI draft as a baseline. Speed + cost reduction.
A native translator edits tone, idiom and UX copy.
Keyword research for each market + meta + hreflang + canonical.
Your TM stays with you. Export TMX anytime.
Choose your pace and scope — always at a fixed price with translation memory. Prices exclude VAT. Amounts shown cover the package scope; additional work requires an agreed scope and a confirmed quote. Small tasks cover 1–2 days of total work, including analysis, communication, QA and agreed revisions. Delivery dates are agreed separately; this allowance is not a delivery deadline commitment.
1–2 days of total work; timing agreed
Delivery in 3-5 weeks
Implementation in 6-10 weeks + ongoing
Every project uses proven tools. AI where it offers an advantage. Humans where tone matters.
AI draft per segment. Custom prompts for each industry / tone. Model chosen for each use case (DeepL for EU languages, Claude for nuance, OpenAI models for creative work).
Native speaker + sector expertise (B2B / e-commerce / legal / medical / technical). AI draft editing for tone, idiom and UX copy.
Meta title + description for each language + alt text + hreflang + canonical. Keyword research for each market (different volume + intent).
Phrase / Lokalise / Crowdin. Reuse segments from previous translations. TMX export (portable when you switch providers).
Buttons, labels, errors, success messages — localised for the market context. Date formats, currency, plurals, gender.
Linguistic QA (proofread native) + functional QA (UI strings fit, character limits, special chars). Bug list per locale.
We review these questions with you before we confirm the scope and deliverables.
Scope and handover
Who will read the material, and what should they do next?
What product information and sources can the writer use?
Who will review the facts, wording and brand voice?
How will we manage revisions, approval and later updates?
The Balance IT standard
Draft from language models or DeepL + native translator editing.
The agreement sets the price, which revisions are included and how we price additional work.
Meta + hreflang + canonical + keyword research for each market.
Phrase / Lokalise / Crowdin. Your TM. Export TMX anytime. Portable when you switch providers.
Each stage has its own team and success metrics. Translation combines AI + human + SEO, rather than relying solely on machine or human translation.
Each stage ends with a demo and approval. Only then do we invoice and move on.
Week 1 · style guide
Weeks 2-3 · production
Week 4 · localisation
Week 5 · final tests
Loading projects...
Answers to common questions about Translation