Platform · Machine Translation

Every Major MT Engine,
One Platform

Wordbee connects to 14 machine translation engines directly and to more than 60 others through a single meta-connector. Your translation memory fills what you have already translated, MT drafts the rest, and every engine runs under your own credentials. Your content is never shared back to train someone else's model.

MT Engine Configuration
14 engines
DeepL
24 pairs
Confidential
Google Translate
108 pairs
Public
Microsoft
76 pairs
Confidential
SYSTRAN
12 pairs
Restricted
Globalese
45 pairs
Confidential
+ 9 more direct engines · 60+ via Intento
Your keys, your rules. Each engine runs under your own API credentials and confidentiality policy.

The Engine Library

Connect the Engine That Fits the Content

There is no single best MT engine. Quality shifts by language pair, domain, and provider, so Wordbee lets you connect as many as you need and pick the right one for each job.

Direct Integrations

Fourteen engines connect out of the box, including DeepL, Google, Microsoft, SYSTRAN, Reverso, Tilde, and Globalese. Each runs under your own API key, so usage and billing stay on your account.

One Connector, 60+ Engines

The Intento meta-connector links more than 60 additional providers and models through a single integration. Connect once, then choose an engine per language pair or let smart routing pick for you.

Engine-Specific Controls

Use what sets each engine apart. DeepL formality, glossaries, and context prompts, or SYSTRAN neural fuzzy adaptation that learns from your translation memory and adapts its output.

Custom and LLM Engines

Run large language model engines such as OpenAI and Mistral with custom prompts for tone and style, or have Wordbee build a connector for an in-house engine that is not on the list.

Routing and Profiles

Send Every Job to the Right Engine

An MT profile is a named connection to one engine with its own settings. Stack several, set their priority, and let routing rules send each language pair where it performs best, always after your memory has had the first pass.

  • One MT profile per engine, domain, or access level
  • Set each profile to enabled, disabled, or editor-suggestions-only (Microsoft, Google, Reverso)
  • Toggle individual source and target languages on or off
  • Route every language pair to the engine that fits it best
  • Directional routing: English to German can differ from German to English
  • Memory always runs first; only the unmatched segments reach MT

From Draft to Delivery

Machine Translation Is a Step, Not a Shortcut

MT in Wordbee is not a standalone button. It sits inside the pipeline next to your memory and quality checks, and every post-edited segment you keep makes the next project start further ahead.

Machine Translation · Pipeline
  1. Every post-edited segment you consolidate raises the match rate on the next project

Step 1 · Configure

Configure the Engine

Connect an engine and enable it in a word count profile. Set the TM match threshold above which MT is skipped, so memory always wins when it has an answer.

At this step

  • Pick the engine per project or per target language
  • Set the TM threshold that MT respects
  • Choose enabled or limited access for the engine

MT Hive

Self-Service Translation for Your Clients

MT Hive is a self-service portal where your clients translate files instantly and, when they need more, escalate to human post-editing in two clicks. It is a separately licensed module inside the Client Portal.

  • Instant translation, no ticket: clients upload or paste, pick a pair and engine, and download a combined TM and MT result
  • Escalate to humans in two clicks: Wordbee creates the order, project, and jobs automatically from a template you preconfigure
  • A form per client, each with its own engine, memory settings, file formats, and DeepL formality, with usage tracked for billing
Client Portal · Acme Corp
4 orders
Product Manual v3.2
6 languages
Delivered
Release Notes Q1
12 languages
In Review
Marketing Brochure
4 languages
Translating
Legal Update
2 languages
Quoted
This Month
24 orders
Avg. Turnaround
1.8 days
Self-service. Clients submit, track, and download without emailing the PM.

Data and Confidentiality

Your Content Stays Yours

Wordbee sends only what an engine needs to translate, and corrected segments are saved to your own memories, never shared back to train the provider's model. Every engine runs under your own keys, on a platform built for confidentiality.

Security Status · Wordbee
Confidential by default
  • Your MT Data

    Runs under your own API keysNot shared back to train the providerSaved only to your own memories
  • Encryption

    AES-256 at restTLS 1.2+ in transit3rd-party penetration testing
  • Data Residency

    EU regionUS regionHosted on · Microsoft Azure
  • Certifications and Access

    ISO 27001 CertifiedGDPR Compliant · signed DPASAML 2.0 SSO + MFA125 scoped access rights

Common Questions

What Teams Ask

Ready to Put Machine Translation to Work?

Start a free trial and configure your engines on your own content, or explore the rest of the platform that runs alongside it.