The Done blog.
How to hand real work to AI, and keep track of it.
· 7 min read
Who approves an AI agent's work? 22 task apps audited
We checked 22 task and project apps: 17 let outside AI agents in over MCP, but none makes a person approve what an agent finishes by default. The data.
Read the postMore posts
· 7 min read
Task app MCP servers audited: auth, tools and approval
We audited the official MCP servers of 14 task and project apps: all use OAuth, 6 publish a tool count, and none holds an agent's work for approval.
· 6 min read
Read vs write: how 4 task app MCP servers split 271 tools
We sorted the 271 MCP tools of Done, Jira, Notion and ClickUp into reads and writes: 52% can change your data. What that means for approvals and annotations.
· 7 min read
Task manager pricing in 2026: 22 apps and their AI bills
What 22 to-do and project apps cost in October 2026: 17 have a free plan, the median paid plan is $9 a month, and 10 now meter AI with credits.
· 5 min read
Connected Done to your AI — now what? Five prompts to try first
You connected Done to Claude or ChatGPT. Five prompts to paste next — add, import, break down, work the queue, schedule — and the MCP tools each one calls.
· 8 min read
The 9 best to-do list apps in 2026
The best to-do list apps in 2026, compared: Done, Todoist, TickTick, Microsoft To Do, Things 3, Any.do, Google Tasks, OmniFocus and Habitica.
· 4 min read
How do I tell an AI agent what "done" means?
Write acceptance criteria as a checklist. How Done's acceptance_criteria and check_criterion give an agent a bar to work against, and why ticking every box still isn't done.
· 4 min read
How do I stop an AI agent that's stuck in a loop?
Pause the task, or the whole board, with a reason. How Done's pause_task and pause_board stop an unattended agent for good, and why only a human can resume.
· 4 min read
What should an AI agent do when it needs a human decision mid-run?
Park the question on the task and move on. How Done's ask_decision tool lets an unattended agent ask you something without stopping its run or guessing.
· 6 min read
Agent lanes: give more than one named AI agent its own queue
Agent lanes give each named AI agent on your Done account its own task queue. How to pin a task to one agent and pass work between agents with handoff_task.
· 6 min read
How to connect Notion, GitHub, and Gmail to an AI agent's tasks
Connect Notion, GitHub, Gmail, Drive, Slack and more to Done so an AI agent can find and read real context before it works a task. No copy-pasting.
· 6 min read
Workspace-scoped MCP tokens: limit an AI agent to one board
A workspace-scoped MCP token confines an AI agent to one shared board instead of your whole account. What it can do, what it can't, and how to set it up.
· 6 min read
Triggering an AI agent instantly: Run now, @claude, and webhooks
Start a connected AI agent on a task right now with Run now, an @claude mention, or a signed webhook that wakes an agent that isn't running yet.
· 8 min read
The 6 best AI to-do list apps in 2026, compared
The best AI to-do list apps in 2026, compared: Done, Todoist, TickTick, Things 3, Motion and Microsoft To Do. What each does best and which are free.
· 6 min read
MCP tool annotations: readOnlyHint, destructiveHint explained
What readOnlyHint, destructiveHint, idempotentHint and openWorldHint mean in MCP, and how Done sets these tool annotations on every tool it exposes.
· 7 min read
Claude task manager: connect Claude to your to-do list over MCP
How to connect Claude (claude.ai, Claude Desktop or Claude Code) to your to-do list in Done over MCP, what it can do there, and how to keep it scoped.
· 6 min read
ChatGPT task manager: connect ChatGPT to your to-do list over MCP
How to connect ChatGPT to Done over MCP, what changes (and what doesn't) compared to Claude, and how the same guardrails apply no matter which model is attached.
· 6 min read
Running a fleet of AI agents on one task board
How to run more than one AI agent against the same task queue without two of them grabbing the same work — atomic claims, live lease status, and auditable runs.
· 6 min read
How to break a big task into subtasks with AI
How to break a big task into subtasks with AI: auto-decompose it, lay the steps out as a plan, and approve each step so an agent can finish the work.
· 5 min read
Recurring tasks and dependencies: a to-do list that runs itself
Recurring tasks save you re-typing the same to-do every week, and dependencies stop an AI agent from starting blocked work. How both work in Done.
· 5 min read
The AI daily digest: a five-minute briefing on your tasks
Done's AI daily digest shows what's overdue, due today, gone quiet, and what an agent shipped, so your morning check on the task queue takes five minutes.
· 6 min read
Goals, projects and tasks: structure work for AI agents
A light three-tier structure of goals, projects and tasks keeps a big initiative readable for you and for an AI agent, without becoming a chore.
· 6 min read
Read-only and project-scoped MCP tokens for AI agents
The trust question isn't which AI model to use — it's how much access it has, how fast you can cut it off, and what still needs a human even after you've connected it.
· 6 min read
AI to-do app vs. AI project management software
Project management software coordinates many people across projects. An AI to-do app helps one person or a small team finish the work. How to choose.
· 5 min read
Turn Slack messages into tasks an AI agent can finish
Capture and update tasks from Slack with /done and @done, then hand the ones you don't want to do to an AI agent — without ever leaving the channel the ask came from.
· 5 min read
What is an AI to-do app? A plain-English guide
An AI to-do app is a task manager that uses AI to capture, organize and even complete your tasks. How it differs from a normal to-do list, and what to look for.
· 5 min read
What is MCP (Model Context Protocol)? A plain-English guide
MCP, the Model Context Protocol, is an open standard that lets AI agents connect to tools and data. What it is in plain English, and why it matters for tasks.
· 6 min read
How to work with AI agents without losing control
A practical playbook for working with AI agents on real work: what to delegate, how to keep a human in the loop, and how to set up guardrails so autonomy stays useful.
· 4 min read
Assigning a task to an AI agent, without losing the plot
How to assign a task to an AI agent in Done: pick the agent, let it do the work and report back in the task, and stay the accountable owner who approves it.
· 4 min read
Put Claude or ChatGPT to work on your task queue
Done is MCP-native, so any AI agent can connect to your account and clear the work you've queued for it. Here's how a run works and how to schedule it.
· 3 min read
Why every AI task needs an approval gate
An approval gate keeps a human in the loop: an AI agent can do the work, but the task can't close until you approve it. How Done's gate works and why.
· 6 min read
How to organize a to-do list that actually gets done
A simple, durable system for organizing a to-do list: capture everything, keep one list, decide what's next, and protect focus. Plus where AI genuinely helps.
· 4 min read
Natural-language task capture: plan your day in one line
Natural-language capture turns a sentence like "email the team friday, urgent" into a structured task — no forms. Here's why it matters and how to use it well.
· 5 min read
When to delegate a task to AI, and when to do it yourself
A clear decision framework for when to delegate a task to an AI agent and when to keep it — based on stakes, reversibility, and how well you can brief it.
Have AI complete your to‑do list.
Free to start. No card.