# DoneThat > DoneThat is remote work's missing piece. It reconstructs your day from desktop activity without timers, then turns that history into summaries you choose to share with your team, goals, coaching, and context an AI assistant can query. Screenshots are never stored: capture is processed in real time and deleted immediately, and only derived structured activity persists. Excluded apps are redacted on the user's own machine before anything is sent for analysis. With a bring-your-own model (Gemini or any OpenAI-compatible endpoint), raw activity never reaches DoneThat's servers at all. The desktop app is open source. Agents can query a user's own DoneThat data over MCP. The Streamable HTTP endpoint is https://mcp.donethat.ai (stateless, OAuth 2.1); it is documented at https://donethat.ai/mcp and described at https://donethat.ai/.well-known/mcp/server-card.json. ## Start here - [DoneThat: Automatic Time Tracking for Remote Work](https://donethat.ai/): DoneThat tracks your work automatically and privately, so remote workers stay focused, in the loop and visible. No timers, no meetings, no spyware. - [Pricing](https://donethat.ai/pricing): DoneThat is free for 14 days. After that, pick Basic, Core or Pro, billed monthly or yearly, depending on the AI models, history and coaching you need. - [Download](https://donethat.ai/download): Download the DoneThat desktop app for Mac, Windows, or Linux to start AI time tracking and automatic productivity summaries. - [FAQ](https://donethat.ai/faq): The most common questions about DoneThat, privacy, pricing, and how the product works. Tell us if something is missing. ## Product - [Privacy](https://donethat.ai/data): Learn how DoneThat protects your data with privacy controls, secure architecture, local capture options, and AI data handling. - [Slack](https://donethat.ai/product/slack): DoneThat posts daily and weekly summaries to a Slack channel automatically, with optional review before they go out. Connect from Social > More > Connect to Slack. - [Integrations](https://donethat.ai/product/integrations): Connect DoneThat to Slack, email, MCP, Zapier, n8n, and the DoneThat API. Share summaries and pull work data into your existing tools. - [Roadmap](https://donethat.ai/roadmap): See the DoneThat product roadmap, upcoming features, and what we are building next for AI time tracking and productivity. - [Install DoneThat on Linux: Setup Guide](https://donethat.ai/install/linux): Install the DoneThat Linux app and get started with AI time tracking, automatic summaries, and productivity coaching. ## Developers and agents - [MCP](https://donethat.ai/mcp): DoneThat MCP server documentation: connect Claude, ChatGPT, and other MCP clients to your work history with OAuth 2.1. - [API](https://donethat.ai/api-reference): DoneThat HTTP API reference: pull aggregated activity, generate summaries, and connect MCP clients programmatically. ## Company - [About](https://donethat.ai/about): Who we are, how to reach us, and where to read privacy, terms, the trust portal, and the DPA, plus links to the founder story and data practices. - [Founder](https://donethat.ai/founder): Founder-led AI consulting on productivity systems, operating design, AI adoption, and workflow improvement. - [Engineers](https://donethat.ai/engineers): Meet the DoneThat engineers who turn measured work insights into shipped automations, with n8n, Claude Code, and the tools your team already runs. - [Community](https://donethat.ai/community): Join the DoneThat Slack community to share productivity hacks, find accountability groups, and give product feedback. - [Support](https://donethat.ai/support): Get DoneThat support, book a demo, troubleshoot the app, or contact the team about setup, billing, and product questions. - [Compare](https://donethat.ai/compare): DoneThat vs time tracking, AI memory, coaching, and productivity tools we think are worth comparing. - [Christoph's Live Profile](https://donethat.ai/live-profile): Meet Christoph Hartmann through consulting and a public DoneThat profile on his own work: AI productivity, impact, and giving back. ## Privacy and legal - [Privacy Policy](https://donethat.ai/privacy): Read the DoneThat Privacy Policy covering data collection, account information, usage data, AI processing, and your privacy rights. - [Terms of Service](https://donethat.ai/terms): Read the DoneThat Terms of Service for account use, subscriptions, payments, acceptable use, and legal terms. - [DPA](https://donethat.ai/dpa): Review the DoneThat Data Processing Agreement for GDPR, controller-processor terms, security measures, and subprocessors. ## Reference - [Reports](https://donethat.ai/api-reference/reports): Pull aggregated activity rows for a date range at activity, day, task, or week resolution. - [Messages](https://donethat.ai/api-reference/messages): Generate formatted day, week, month, quarter, or year summaries - the same content shown on the Summaries tab. - [Projects](https://donethat.ai/api-reference/projects): List, read, create, update, and archive DoneThat projects through the REST API. - [Search](https://donethat.ai/api-reference/search): Search your DoneThat task and activity history. - [Project (deprecated)](https://donethat.ai/api-reference/project): Deprecated operation-based project endpoint retained for existing integrations. - [User](https://donethat.ai/api-reference/user): Read basic profile information for the authenticated API key owner. ## Recent writing - [Async Work Only Works If People Can See Progress](https://donethat.ai/blog/async-work-needs-visible-progress): Async work is not fewer Zoom calls. It fails when the only shared artifact is a calendar or a standup. The teams that make it work keep a cheap, durable progress signal: writing, hill charts, Friday shipped notes. - [Trace the Work. Calendars Are Not a Log.](https://donethat.ai/blog/calendars-are-not-a-work-log): Rahul Vohra found 6-7% of his week on craft work a calendar would never have shown. A calendar is intention. A switch log is the trail. Microsoft telemetry explains the gap, and memory is not a backup. - [What Google Snippets Actually Were (and Why Written Updates Beat Status Meetings)](https://donethat.ai/blog/google-snippets-written-status-updates): Google Snippets were not search-result snippets. They were a Monday email asking what you did last week and what you plan next. Here is how the ritual worked, why status meetings fail, and how writing replaces the standup. - [How to Convince Your Boss to Let You Work Remotely](https://donethat.ai/blog/how-to-convince-your-boss-to-let-you-work-remotely): Ideology loses. A reversible mid-week trial with a paper trail of output wins. What bosses fear in 2026, the Trip.com experiment that updated their beliefs, and a seven-step playbook with no fake sick day. - [Remote Trust Is a Signal Problem, Not a Surveillance Problem](https://donethat.ai/blog/remote-trust-is-a-signal-problem): Leaders lose remote-work trust when hallway context disappears, then reach for RTO or bossware. The evidence points to a signal gap. Fix visibility with outcomes, writing, and reconstructed work. - [Why You Cannot Categorize Time From Memory](https://donethat.ai/blog/you-cannot-categorize-time-from-memory): Reconstructing how you spent a week from memory or a calendar fails. Sam Corcos logged 17,784 hours and still would have put recruiting at 20% instead of 3%. Contemporaneous logs plus a small category set is the practice. - [You Can't Measure Outcomes. Measure Time on Goals Instead.](https://donethat.ai/blog/measure-time-on-goals): Everyone says measure outcomes. For individual knowledge work that mostly doesn't work — outcomes lag, they can't be attributed, and luck dominates. Here is the measurable alternative: time on goals, paired with energy. - [Dogfooding DoneThat: AI Watched My Screen for A Year](https://donethat.ai/blog/dogfooding-donethat): What happens when an AI tracks your screen every 5 minutes for a year? Data-driven productivity insights from 1,585 hours of solofounder work. - [You should probably track your time (and it just got easier)](https://donethat.ai/blog/you-should-probably-track-your-time): Most people in EA know they should track their time but don't, because existing tools are either too much work or not accurate enough. Here's why it's worth doing and a new AI-powered option that fixes both. ## Optional - [Blog index](https://donethat.ai/blog): Every post, newest first. - [AI adoption knowledge base](https://donethat.ai/ai-adoption): A large library of AI adoption use cases by domain and stage. Enumerated in sitemap.xml rather than here. - [Full sitemap](https://donethat.ai/sitemap.xml): Every indexable URL. - [Blog feed](https://donethat.ai/blog/rss.xml): RSS 2.0.