ModelFront today announced the launch of OpenLoc, the open TMS built for AI at Tech Week in San Francisco. OpenLoc is free, open, AI-native enterprise translation management software designed to help localization teams thrive in the AI era by removing AI taxes, per-seat and per-word pricing, and closed integrations that drive shadow localization.
The TMS creates value in the AI era: reusing human-reviewed content, managing terminology and style, handling and validating formatting, systematic evaluation, vendors and billing, and dashboards.
But traditional TMS software was built before AI. Pricing is per seat and per "processed" word, designed for human translation. AI is sold on top: an add-on, a tier, a quota, and often teams cannot bring their own model. Managing their own prompts is the exception, not the rule, and giving an LLM proper context is hard. Integrations with AI, content and analytics often cost extra. This friction causes “shadow localization," whereby the teams across the enterprise with content to translate skip enterprise translation altogether.
With OpenLoc, localization teams can run raw MT and AI workflows with no AI tax, use any LLM with context and RAG through an easy, powerful prompt builder, get free integrations similar to installing an app, create dashboards from chat or in other tools, manage anything via system-wide MCP, and host themselves for free or get enterprise hosting and support at cost.
“In the AI era, open is winning more than ever. Free, open, AI-native enterprise translation management is key for localization teams to not just survive but thrive,” said Adam Bittlingmayer, CEO and co-founder of ModelFront.
The full OpenLoc source code is available under the Business Source License 1.1 by default, which converts to an open-source license. Other licenses can also be requested. The source code will be released publicly on GitHub on October 19, 2026. Open-source has been the tried and true path to progress in software for decades. Openness goes beyond reading or running source code, or reducing cost. Open allows teams to get more features sooner. OpenLoc is the open harness for the enterprise translation vertical.
Enterprise localization teams can host OpenLoc themselves, or get enterprise hosting and support provided by ModelFront at cost, or by third parties. Migration between systems is getting easier thanks to MCP and free integrations. Traditional TMSes increasingly support MCP, so an agent can read not just translation memories but workflows and settings, and rebuild them in OpenLoc. And with OpenLoc, anything can be done via MCP, so teams can migrate away at any time, workflows and settings included.
ModelFront today also announced the OpenLoc Early Access Program: everything a team needs to switch, for free. Selected companies get enterprise hosting and support, AI credits, integrations and features built for them, and help with their migration. The ModelFront team will be at Vancouver Localization Week to demo OpenLoc and answer questions (LocWorld booth 46) on October 19–21.
About OpenLoc
OpenLoc is supported by ModelFront, which provides AI to check and fix AI translations while keeping human quality.
“Think of it like Google supporting Chrome and Android. You're free to use any search engine, so Google has to keep being the best search engine. Likewise, you're free to use any AI for QE and APE in OpenLoc, so ModelFront has to keep being the best AI for QE and APE — it's our main and only business,” said Bittlingmayer.
The launch of OpenLoc was announced at Tech Week 2026. The first-ever translation AI event at Tech Week featured speakers from Google, Microsoft, DeepL, ModelFront, Krisp, Meta, Adobe, Disney, Twitch and LocWorld.
About ModelFront
ModelFront provides AI to check and fix AI translations while keeping human quality. ModelFront uses outcome-based pricing, so it is aligned with better AI outcomes. Learn more at modelfront.com.
About translation
Generative AI started in translation – the Transformer model architecture, the foundation of large language models, was invented for machine translation at Google in 2017.
Translation was thus the first workflow to shift to humans verifying and editing AI output, rather than manually writing content from scratch.
Translation is also key to human progress. More than 6 billion people cannot understand English, and companies spend $70 billion a year, and growing, on translation and localization.
