February Product Update: Trusted AI, Now Generally Available
February was a meaningful step forward for teams moving from AI experimentation to production deployment. We advanced the Enterprise Context Engine to general availability, expanded governance controls, and introduced deeper reporting and automation capabilities across Agent, CLI, and Admin experiences.
Here’s what matters most.
Enterprise Context Engine Is Now Generally Available
The Enterprise Context Engine is now GA across Enterprise SaaS and self-hosted environments.
This milestone is foundational. The Context Engine builds a structured, living model of your repositories and services so AI can reason about real dependencies, APIs, and architectural relationships. The result is fewer blind suggestions, safer automation, and AI that operates with awareness of your actual systems.
Now available:
- Service summaries per repository
- OpenAPI specifications per repository
- Organization-level repository definitions
- A new Admin Settings section to control model selection, scheduling, availability, and exposed tools
- An Assets view to inspect generated contextual artifacts
If you are serious about Trusted AI in production environments, structured context is no longer optional. It is the baseline.
Deeper Agent Reporting and Usage Visibility
Scaling AI requires visibility. This release introduces expanded reporting to make adoption measurable and auditable.
New capabilities include:
- Enhanced usage and performance metrics in Enterprise reports and APIs
- A dedicated Agent Usage UI for viewing activity and consumption
AI without reporting creates governance blind spots. These updates help teams understand impact, monitor adoption, and manage risk with confidence.
Expanded Governance and Enterprise Controls
Trusted AI requires both flexibility and enforcement. February introduced several new enterprise controls:
- User-level controls to enable or disable MCP tools
- Personal Access Tokens (PATs) for secure authentication of scripts, integrations, and APIs
- Bring Your Own AI support, allowing enterprises to integrate their preferred LLM providers while preserving Tabnine governance and experience
These capabilities give organizations tighter policy control while maintaining architectural flexibility.
Smarter Agent Capabilities in the IDE
We also enhanced day-to-day developer workflows with improvements designed for performance and precision:
- ripgrep-powered search for fast, code-aware exploration
- Offset and limit mechanisms for reading large files without bloating context
- A new /code-review slash command
- Image context support for Gemini models to reason over screenshots and diagrams
- Visual Studio 2026 support
On the model and platform side:
- Added support for Claude Opus 4.6
- Expanded support for additional self-hosted models
These updates strengthen agent reliability while expanding model choice across enterprise environments.
Upcoming Webinar: Trusted AI in the Enterprise
With the Enterprise Context Engine now generally available, we are hosting a dedicated webinar focused on operationalizing Trusted AI in real engineering organizations.
We will cover:
- Why AI without structured context fails in complex environments
- How context-aware systems reduce blast radius and rework
- What governance looks like when scaling AI across teams
- Practical architecture considerations for enterprise deployment
If you are evaluating how to scale AI safely across your engineering organization, this is the moment to engage. Register here.
February was about moving from promising AI to production-grade AI. Trusted AI is no longer a roadmap item. It is now generally available.
