Cursor Adds Context Usage Breakdown for AI Agents

TL;DR

• Cursor has released version 3.3, introducing a new feature for AI agents.
• Users can now view a breakdown of their AI agent's context usage.
• This allows for diagnosing context issues and optimizing agent setups.

The latest update to Cursor, version 3.3, released on May 6, 2026, brings a valuable new tool for developers working with AI agents. This significant enhancement allows users to gain unprecedented visibility into how their agents are utilizing context. For anyone building or refining complex AI workflows, understanding context consumption is crucial for performance and efficiency.

This new context usage breakdown feature empowers developers to scrutinize the data sources and information their agents are accessing. Whether it's through rules, skills, MCPs, or subagents, the granular statistics provided will shed light on where an agent is drawing its knowledge from. This level of detail is essential for identifying inefficiencies or unexpected behaviors within an agent's operation.

By providing these diagnostic tools, Cursor aims to help users optimize their agent setups. Developers can now pinpoint exactly which parts of the context are being heavily relied upon, and conversely, which might be underutilized or causing performance bottlenecks. This information is instrumental in fine-tuning prompts, refining agent logic, and ultimately building more robust and responsive AI assistants. For more in-depth information on how agents handle context, you can refer to the Agent Prompting and Context Usage documentation.

The introduction of this feature in version 3.3 marks a step forward in making AI agent development more transparent and manageable. It caters to both seasoned AI engineers looking to squeeze every bit of performance out of their agents and newer users seeking to understand the inner workings of their AI tools. Developers can keep track of all new features and updates by visiting the Cursor Changelog.

Summary

• Cursor version 3.3 introduces a context usage breakdown for AI agents.
• This feature aids in diagnosing and optimizing agent configurations by revealing context consumption patterns.
• Developers can gain better insights into their agent's data sources and improve overall performance.

Source: Context Usage Breakdown · Cursor