DoneThat

What DoneThat does

DoneThat MCP server is presented as a privacy-first system for tracking work and making that context available to AI agents. Its stated capabilities cover four areas: summaries, reports, coaching, and AI-ready long-term memory. Together, these capabilities describe a service intended to help assistants understand work activity, produce condensed views of it, and retain useful context across interactions.

The listing does not define the underlying work-tracking model. It does not say whether tracking is manual, automatic, event-based, or connected to a particular project-management, calendar, communication, or source-control product. Developers should therefore evaluate DoneThat as a work-context and memory service, without assuming support for any external integration.

How it works

The available description identifies DoneThat as an MCP server, so its purpose is to make the product’s work-tracking context available to an AI agent through the Model Context Protocol. The supplied material does not describe request formats, returned data, authentication, transport, or whether the server reads, writes, or both.

At the capability level, the DoneThat MCP server is associated with summaries and reports for reviewing tracked work, coaching for guidance based on that context, and AI-ready long-term memory for retaining information that an assistant can use later. These labels indicate the intended categories of assistance, but they do not specify separate callable tools, parameters, permissions, or output schemas.

Limitations and notes

The source material is limited to a short product description, and no README is available. As a result, several implementation details remain undocumented:

• No supported MCP clients are named.
• No installation command or transport is provided.
• No environment variables or configuration keys are listed.
• No authentication method is described.
• No tool names, input fields, output formats, or error behavior are documented.
• No external services or integrations are identified.
• No license is stated.

The privacy-first positioning is part of the product description, but the available material does not explain retention, encryption, data deletion, access controls, hosting, or user-configurable privacy settings. Confirm those details before deploying the DoneThat MCP server with sensitive work information. The same caution applies to the long-term-memory capability: the description establishes its intended use, but not what data is retained, for how long, or how memory is managed.

Based on the available information, DoneThat is most relevant when an agent needs persistent work context, generated summaries or reports, and coaching-oriented assistance. A technical evaluation will require additional documentation covering setup, security, supported clients, and the exact MCP interface.