Contextberg icon

Contextberg

Contextberg records your screen, input, browser, and AI agent history locally, then shares work context via MCP with Claude Code and Cursor for resumptions, search, and reviews.

Contextberg

An app that connects local work history to AI agents

Contextberg is a desktop work-memory app that records PC screen activity, input, browser history, and AI agent usage history locally. It accumulates the flow of your work in the background, allowing you to review it on a timeline or organize it into daily and long-term context that AI can read.

Recorded information can be provided through the built-in MCP server to supported agents such as Claude Desktop, Claude Code, and Cursor. It is useful for having the agent summarize your recent situation after an interruption or search materials you viewed and error logs from the past. Data is primarily stored locally, and the Free plan supports local storage and Local LLMs.

Key features

Automatic work context recording

Records screen activity, windows, apps, input, browser history, and usage history from Claude Code, Cursor, terminals, and other tools in the background.

Timeline and usage review

Review chronological screenshots and usage time by app on a timeline to look back on the day’s workflow.

Three types of memory generation

Organizes work information into three types—activity memory, daily memory, and long-term memory—to generate daily reports and context that agents can use.

Agent integration via MCP

Uses the built-in MCP server to provide recorded context to supported agents such as Claude Desktop, Claude Code, and Cursor.

Summaries for resuming work

Summarizes work before you stepped away by combining recent activity, browser history, and agent history. You can explore the summary further in chat.

Inference destination selection

Choose where data is stored and where inference takes place. In addition to local models and LM Studio, it supports Codex sign-in, Gemini, OpenRouter, and Contextberg Cloud.

Use cases

  • Resuming interrupted development work

    After stepping away or switching to another task, review a summary spanning your recent screen, browser, and agent history, then resume work without repeating the explanation.

  • Searching past research and troubleshooting

    Have an AI agent search materials you viewed and error logs from the past, bringing previous research and solutions back into the conversation.

  • Reviewing personal work habits

    Use the timeline, usage time by app, and daily reports to reflect on when you were focused and when you drifted off task unconsciously.

  • Daily reviews and accumulating records

    Have AI summarize the day’s work log and accumulate it in your personal work database as material for daily reports and progress reviews.

  • A private AI work environment

    Choose a local model or LM Studio to handle work history primarily locally while providing AI agents with the context they need.

Pros and Cons

Pros

  • You can get started without registering an account, and it supports Windows and macOS.
  • It collects screen, browser, input, and agent history separately and organizes them into a timeline and memories.
  • The MCP server is built in, allowing work context to be connected to supported AI agents.
  • You can choose from multiple inference destinations, including local models and LM Studio.
  • A Free plan is available at no cost indefinitely, letting you try basic recording, analysis, and MCP integration.

Cons

  • Because it records a wide range of work information, including screen activity and input, you should review the recording targets and privacy settings before use.
  • AI models, data retention periods, and the scope of Gemini access vary by pricing plan.
  • Supported operating systems are macOS 13 and later and Windows 10/11. Supported agents and specific integration settings must be verified for your environment.

FAQ

Is an account required for setup?

Contextberg is downloaded from the official website and launched using an installer. No account registration or complex environment setup is required, according to the product information. The Windows version is distributed through the Microsoft Store, while the macOS version is provided as a DMG.

Which operating systems are supported?

It supports macOS 13 and later, as well as Windows 10/11. The Windows version is available from the Microsoft Store, and the macOS version can be obtained as a DMG from the official website.

What types of work data can it record?

When the app runs in the background, it records windows and apps, input, the screen, browser history, and agent usage history from Claude Code, Cursor, terminals, and other tools. The timeline shows chronological screenshots and usage time by app.

How does it connect to AI agents?

Enable the built-in MCP server and add Contextberg to a supported agent such as Claude Desktop or Cursor. You can then search past work logs or give instructions such as “Summarize the material I was looking at earlier.”

What are the differences between the pricing plans?

The Free plan is free forever and includes local storage, AI work analysis, Local LLM support, and MCP integration. Lite costs $8 per month and provides access to all Gemini models and unlimited data retention, while Pro is described as in development. The Free plan includes only the basic Gemini model, with data retained for 180 days.

Quick Facts

Category
Productivity / Developer tools
Supported operating systems
macOS 13 and later, Windows 10/11
Primary users
Individual users and developers who use AI agents
Integration method
Built-in MCP server
Free plan
Free forever, 180-day data retention
Lite plan
$8/month, all Gemini models and unlimited data retention