🆕 Highlights
Introduction of MCP, enabling connection between Trebellar’s data model and external AI systems
Expanded asset and location APIs, along with new asset management tooling
Portfolio Planner improvements across forecasting, grouping, and prediction accuracy
Continued Agents infrastructure improvements and Intercom integration
Ongoing infrastructure upgrades, including workflow migrations and service updates
Continued stability improvements across planner, asset history, geo handling, and caching
📦 Module-Based Breakdown
Platform & MCP
March introduces support for MCP (Model Context Protocol), enabling Trebellar’s structured data model and ontology to connect with external AI systems such as ChatGPT, Claude, Gemini, and Manus.
This establishes a foundation for using Trebellar data within broader AI workflows, allowing external tools to query and interact with portfolio, planning, and asset data in a structured and consistent way.
Additional updates expand asset and location APIs, along with new asset management tooling to support more robust data modeling and interaction across the platform.
Portfolio Planner
Planner continued to improve across forecasting accuracy and usability. Updates include:
Improved prediction accuracy, including fixes to p90 forecasts
Support for comparing simulated and ML-predicted attendance
Improved grouping logic and navigation behavior
Fixes to calculation issues, including area aggregation
These updates improve confidence in forecasting outputs and support more flexible scenario modeling.
Agents & Integrations
Work continued on strengthening the Agents foundation, including improvements to response handling and stability.
Intercom integration was introduced and refined, supporting improved communication and support workflows directly within the platform.
Data, APIs & Asset Management
Enhancements to asset and location APIs improve data consistency and expand support for asset management workflows.
Additional fixes address asset history retrieval, coordinate handling, external ID consistency, and caching behavior.
Infrastructure & Stability
March included continued investment in infrastructure and platform reliability. Updates include workflow migrations, Node upgrades, service updates, and improvements to CI processes.
Ongoing fixes across planner, geo handling, and caching improve overall system performance and stability.
✅ Summary
The March 2026 release introduces MCP connectivity, expanding how Trebellar data can be used within external AI systems, while continuing to improve Planner accuracy and overall platform reliability.