One open, AI-native platform for Red Hat's global content.
Red Hat is building open, enterprise AI, from Granite to OpenShift AI. LILT is the same for your content: one open, agentic AI platform that is model-agnostic, so you can orchestrate Granite or any model, and self-managed or air-gapped to run in your hybrid cloud. Instant AI localizes technical docs, developer, and community content at open-source scale, with expert verification only where it matters, and volume economics that keep it affordable.
Use cases
How Red Hat wins with LILT
Technical documentation at scale
RHEL, OpenShift, and Ansible docs localized across languages, fast enough for the release cadence.
Developer and community content
Localize developer guides, tutorials, and community content so contributors everywhere work in their own language.
Instant AI, verified where it matters
Route high-volume content to instant AI translation, and add agent or expert verification only where it counts, on volume economics.
Model-agnostic and open
Orchestrate Granite, your own models, or third-party LLMs, with no lock-in, aligned with your open strategy.
One source of truth
Consolidate a multi-vendor stack (XTM, Phrase, and in-market agencies) onto one governed platform, with one glossary and a single view of cost, quality, and turnaround.
Agent-native, in your hybrid cloud
A localization agent that fits an MCP-native, hybrid and OpenShift world, self-managed or air-gapped.
Enterprise trust, built in
Your content and data stay yours
Custom models trained only on your content. Your data is excluded from anyone else's model training, with automated PII redaction and role-based access.
Self-managed, air-gapped, and sovereign-ready
Run in your own hybrid cloud or fully air-gapped for regulated and sovereign content, meeting strict data-residency and data-control requirements.
One governed system
One glossary and style guide, risk-tiered human review, and real-time MQM analytics with a full audit trail.
Open and model-agnostic
Bring your own LLM and orchestrate any model, with no black box and no lock-in.
One platform, not a patchwork
Red Hat's teams have run a multi-TMS stack (XTM and Phrase). LILT connects generative AI to all of your systems, guarantees quality and consistency, and adapts in real time, so you scale on one open platform instead of stitching tools together.
Contextual AI
First-party managed services
Reach
Model training
Content generation
Quality assurance
Built for how Red Hat works
Your teams do not need another silo. LILT connects to the systems and pipelines your teams already use.
100+ connectors, API and CI/CD
Connect your content systems and code repositories, and trigger translation inside your build and docs pipelines. Around 80% of LILT work flows through connectors, not files.
Model-agnostic orchestration
Orchestrate Granite, your own models, or third-party LLMs in one platform.
Agent-native via LILT MCP
A localization agent for your MCP-native, hybrid pipelines, with brand governance built in.
