Niches · Translation

Translate text that must stay on own servers

No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.

0companies, full fit
4companies, fit with a limit
17companies carry the tag
7share of the category's work
4first-hand accounts

What counts as done

Translation runs on the operator's own server or under documented no-retention terms, has the same languages as the hosted service, and needs no outbound call.

How agents ask for it: translate this offline, translate without sending it to a cloud service, run translation inside our network

Does the job, with a limit 4

It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.

ListingWhy
Azure Translator BB
Microsoft Corporation
Translator containers run on premises; connected ones send only usage metadata, and disconnected (offline) ones need an approved commitment-tier licence paid up front. Language parity with the cloud was not stated on the page. source
Lara Translate API B
Translated s.r.l.
Hosted only, with no self-hosted engine; a noTrace option stops texts being stored for model improvement, which is otherwise the default. source
LibreTranslate D
LibreTranslate
Self-hosts under AGPL-3.0 with no outbound call needed, but an open report found Thai in the pip install and missing from the Docker image, and Argos models are weaker on short mixed-language strings. source
Lingvanex Translation API F
Nordicwise Limited
Lingvanex sells on-premise translation servers and offline SDKs, but the language parity with the cloud API and install details are vendor claims on a product page, not docs. source

Couldn't check 1

The docs were closed to us, or silent. These carry no mark either way.

ListingWhy
SYSTRAN Translate API E
SYSTRAN SAS
The listing names an on-premises SYSTRAN server with the same API, but no install or language-parity docs were read.

Doesn't do this job 11

Listings in the category that do something else, with the reason.

ListingWhy
Amazon Translate BB
Amazon Web Services, Inc.
Hosted only. The guidelines say Amazon Translate might store data to improve models unless the account asks AWS Support or sets an opt-out; there is no self-hosted build. source
Crowdin BB
Crowdin OÜ
Hosted platform only; no self-hosted build was found.
DeepL API BB
DeepL SE
Hosted only, with no self-hosted engine in the docs; data handling for API Developer keys is not stated in the privacy policy. source
Google Cloud Translation BB
Google LLC
Hosted only; a first-hand account moved to self-hosted NLLB to cut the Google API bill. source
Lingo.dev B
Replexica, Inc.
Hosted platform calling third-party LLM providers; no self-hosted option.
Localazy C
Localazy s.r.o.
Hosted platform only.
Lokalise BB
Lokalise, Inc.
Hosted platform only.
Phrase B
Phrase a.s.
Hosted only, and the terms of 4 October 2026 allow training non-text models on customer content.
POEditor MCP server E
Code Whale Inc.
Hosted MCP server calling hosted engines.
Smartcat D
Smartcat Platform Inc.
Hosted platform only.
TranslatePlus E
PetaBytes, Inc.
Hosted API only.

What stops agents finishing

  • self-hosted image lacks languages the hosted version has
  • open models weaker on short or mixed-language strings

Around it

Before: mailbox-access.read-message, file-storage.get-file. After: translation.translate-text.

Where this comes from

Weight 7 of 100 in Translation: 4 of 55 accounts (offline LibreTranslate use, a game mod bundling it, Docker language gaps, on-prem NLLB); no account discussed retention terms directly

The same as data: /api/v1/use-cases/translation.private-translation.json, or the MCP tool get_use_case with id translation.private-translation. Other jobs in this category: /api/v1/niches?category=translation.

Draft, read 2026-11-15 from public docs and first-hand accounts, not yet reviewed by a person. A mark says what the docs show, not what a run measured.

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.