Weaving tasks since 2012.

What's next?

Tapestry starts like a plain todo list. Add links when one thing depends on another, and your list turns into the work you can actually do now.

Achieve happiness

Personal

Toast some bread

Personal

Assemble sandwich

Personal

Buy cheese

Personal

It's a to do list.

It works just like how you'd think.

Personal Call the dentist
Call the dentist

Come back at the right time.

Snooze entire projects at once. Automatically focus tasks if their dependents come due.

Send birthday card
Snoozed until Friday

Plain list

Treat it like a regular todo list.

Type the thing. Press enter. You can add notes, dates, links, and workspace cleanup later.

Personal

Timing

Come back when the work matters.

Snooze a whole cluster when it is not relevant yet. If a dependent task is due, Tapestry pulls the prerequisite chain back into focus.

Snoozing a task also snoozes tasks that depend on it
Adding a due-date prioritizes all tasks that lead to it
"Crunch" views help show you when deadlines get too tight
Not relevant yet
Dependent comes due

Conflict Detection

Dates should agree with the graph.

If A has to happen before B, but A is due after B is due, Tapestry can flag the contradiction. Can your calendar do that?

A
Due Mar 8
B
Due Mar 4
Flagged A depends before B, but A is due later.

Agents

Ready for agentic productivity.

Tapestry is built so humans and agents can work from the same task graph, not two different systems duct-taped together.

Make a plan to book my trip to Italy.
Using Tapestry MCP
Done. I found your existing tasks “book flights” and “visit my bud in Pisa,” then attached the attraction-booking tasks after flights and put “choose dates” before the whole trip chain.
Two tasks are ready now: “choose dates” and “see which museums are must-sees.” I spun up a delegate to research museums, and I self-assigned “choose dates” so we can handle that together. When are you thinking of going?
Delegates

Give agents the keys you actually mean to give them.

Create scoped delegates with read or write access. They can use the same API surface you use, but only inside the boundaries you choose.

MCP + Skill

Let long-running agents plan in the graph.

Agents can break goals into small tasks, attach dependencies, claim work, read comments, and keep walking the DAG as things unblock.

Open the Tapestry MCP client →
Audit trail

Know who did what, even when “who” is a bot.

Delegate names show up beside edits, comments, assignments, and completions, so agent work is inspectable and traceable back to a human agent owner.