locAGI

locAGI suits enterprise localisation and linguistic teams that want translation memory, quality review and AI options in a managed cloud workflow. Its combination of CAT tools and configurable AI is compelling, but the private beta and cloud-only deployment make it a poor fit for teams that need a mature, self-hosted system today.

Overview

Translation memory, terminology management, QA checks, alignment and MQM review bring core localisation work into one platform. Support for established CAT file workflows can help teams exchange projects without committing to an immediate replacement of their existing tools.

locAGI is operated by Terratra OÜ in Tallinn, Estonia, and is currently in private beta, with public release scheduled for Q2 2027. It is a managed cloud service, not a self-hosted or offline desktop tool. That trade-off favors teams comfortable with a web-based workflow and rules out organisations that require local deployment.

For a broader category comparison, see Computer-Assisted Translation Software.

Key features

Translation and quality workflow

Translation memory and terminology management support repeat projects and consistent language choices; QA checks and an alignment tool cover preparation and review. MQM review can use MQM Core, LISA QA Model 3.1, or custom free-text categories, and findings can be exported to Excel. That flexibility suits teams with established review criteria, while the export option gives reviewers a way to carry findings into spreadsheet-based processes.

AI choices and controls

The platform lists models from OpenAI, Anthropic, Google, Mistral, DeepSeek, xAI and DeepL, as well as private models through Ollama or OpenAI-compatible endpoints. Teams can use locAGI-provisioned tokens or BYOK, giving them a choice between a managed AI spend and bringing their own credentials. AI token rates are not published, so the pay-as-you-go cost cannot be judged in advance.

Project audit logs, per-user rate limiting on AI surfaces, BYOK and private-model routing are available controls. SSO, SCIM provisioning, data-residency controls and SOC 2 Type II readiness are roadmap work, with SOC 2 readiness underway; organisations that require those controls now should look elsewhere.

File exchange and automation

Workflows cover memoQ MQXLIFF, Trados SDLXLIFF, Phrase MXLIFF, and XLIFF 1.2 and 2.0. More than 40 source formats are supported, including DOCX, XLSX, PPTX, IDML, HTML, XML, DITA, JSON, PO, RESX and SRT. A REST API covers projects, segments, QA runs and exports, alongside API tokens, webhooks and an OpenAPI specification. These are useful foundations for teams automating localisation work, though the platform cautions that desktop CAT tools are more battle-tested for XLIFF files with 50,000 or more segments.

Pricing

Free forever — platform: free. Platform, seats and setup are billed at $0, with no annual commitment and unlimited seats. Included storage is 1 GB on the homepage; AI is billed separately or can use BYOK. This makes the platform plan accessible for teams that can keep within the storage allowance and manage AI costs separately.

Pay-as-you-go AI: custom pricing, billed by tokens consumed with no monthly minimum. It keeps AI spending tied to usage rather than a monthly minimum, but teams should account for the unknown unit cost before relying on it at scale.

Platforms

locAGI is web-deployed, with API, Linux and macOS platforms also listed. It has no offline desktop use and cannot be self-hosted, so it does not suit teams whose requirements depend on a local installation or disconnected work.

Who it's for

The strongest fit is an enterprise localisation or linguistic team, including internal staff and contracted freelancers, seeking shared translation, QA and AI workflows without a per-seat platform charge. The unlimited-seat free platform plan may be especially useful for larger groups, provided 1 GB of included storage is sufficient and AI is budgeted separately.

It is a weaker fit for production-critical teams unwilling to work in a private beta, organisations that require self-hosting or roadmap compliance controls, and projects with very large XLIFF files that call for battle-tested desktop CAT tools. locAGI says customer suggestions that benefit the platform are implemented free of charge, typically within a week, which may appeal to teams looking to shape a developing product.

Pros and cons

• Pros: Translation memory, terminology, alignment and several QA approaches are brought together, including exportable MQM findings.
• Pros: A broad choice of hosted and private AI models, with BYOK as an alternative to locAGI-provisioned tokens.
• Pros: Unlimited seats and no platform charge lower the barrier to including staff and freelancers.
• Cons: Private beta status makes it a riskier choice for workflows that need a mature production service.
• Cons: Managed-cloud-only deployment excludes self-hosting and offline desktop work.
• Cons: Desktop CAT tools are more battle-tested for XLIFF files at 50,000 or more segments.

Alternatives

BasicCAT is the clearer choice for readers who want a free, open-source CAT tool with Linux, macOS and Windows support rather than a managed cloud platform.

Supervertaler Workbench is another free option for readers seeking a CAT tool on Linux, macOS or Windows.

TextUnited may fit a small team that wants a defined monthly allowance: its free plan includes 15,000 words, 15 PDF pages and one user each month, while its paid team plan costs €150 a month.

CafeTran Espresso suits readers who prefer a desktop CAT tool across Windows, macOS and Linux, with a free tier capped at 1,000 translation units and 500 glossary terms or an annual subscription at €80 a year.

XTM Cloud is an alternative for teams considering a paid cloud platform; its Team plan costs $16,500 a year and includes 2 million processed words per year and unlimited users.

Déjà Vu X3 offers a Windows free plan limited to one target language per project and one translation memory and termbase per target language, making it an option for readers who can work within those limits.

Fluency Now is a paid alternative for Linux, macOS and Windows users, with its Professional plan at $15 a month for up to two computers and a yearly payment option with two months free.

Swordfish is another paid desktop option for Linux, macOS and Windows; its annual subscription is $140 a year, and the subscription key cannot be shared or transferred between machines.

Verdict

Choose locAGI if your localisation team wants a shared cloud CAT workflow with flexible AI-model choices, configurable quality review and no platform charge per seat. Look elsewhere if you need a proven production platform now, self-hosting, offline desktop work or compliance controls beyond the current roadmap.