Project Blueprint
Describe a goal in a sentence and Didi drafts the whole project: tasks, subtasks, priorities, and due dates. You edit everything before it saves.
By ToDidIt Product TeamUpdated
Project Blueprint is the part of ToDidIt that turns one sentence into a working project. You describe the outcome you are after, Didi drafts the tasks, subtasks, priorities, and due dates that get you there, and you edit the whole thing before a single row is saved.
The problem it solves is the blank project. Everyone knows the first twenty tasks of a product launch or an office move, and writing them out by hand is forty minutes of typing things you already knew. Blueprint does the typing. The judgment stays yours, which is why nothing commits to your ToDidIt workspace until you press Save & start.
- Input: one or two sentences describing a goal.
- Output: a full project tree, up to three levels deep, with priorities and relative due dates already set.
- Cost: 25 To-Dids per generation, 8 per refinement.
- Plan: Squad and up, plus three generations inside the 14-day trial.
- Commitment: none until you save. The draft lives in the browser, not the database.
Go to Blueprint, describe what you are trying to accomplish in a sentence or two, and Didi drafts a full project tree: tasks, subtasks, priorities, and relative due dates. Usually under sixty seconds from prompt to a real project.
Ask for a cadence and you get one. A prompt that says "set the full audit to repeat quarterly" comes back with that task already repeating, on the task that represents finishing the work rather than on all twenty. Nothing else repeats unless you asked, because a project where every task respawns is unusable.
The model returns dates as offsets rather than calendar days, and ToDidIt converts them against your profile timezone before showing them. That is why a fresh draft reads "Due Tue May 12" instead of "in 5 days", and why the same Blueprint generated in Auckland and in Denver lands on sensible local dates in both.
Nothing is saved until you say so.
Blueprint is at its best on goal-shaped work: a defined outcome, a sequence of steps, and enough conventional structure that a good draft is recognisable the moment you read it. Four shapes come up constantly.
- The agency retainer kickoff. "Onboard a new SEO client through the first 90 days." You run the same shape every time and rebuild it every time. Blueprint drafts it, then save it as a Project Template so the next client costs nothing.
- The product launch. "Launch the iOS app to the App Store in six weeks." The dependency-heavy middle (assets, review, staging, comms) is exactly the part people forget under deadline, and it is the part a draft is best at surfacing.
- The personal move. "Move house across the country in two months." Nobody has a template for this because nobody does it twice in a row. That is precisely when a first draft beats a blank page.
- The course or content build. "Write and record an eight-module onboarding course." Repetitive per-module structure that is tedious to type and trivial to generate, then trim.
Tip
The more the prompt reads like a commitment, the better the draft. "Launch the iOS app to the App Store by March 14 with a beta group of 50" produces a usable plan. "App stuff" produces a shrug.
Roughly five minutes end to end, and it costs 25 To-Dids. If you are inside the Squad trial, this is one of your three generations, so spend it on a project you actually intend to run.
Open Blueprint.
Open BlueprintIt is a full-page workspace, not a modal, because you are going to be reading and editing a tree rather than filling in a form.
Check which workspace you are in.
The draft saves into the active workspace. If you run client work and personal work side by side, switch before you generate rather than moving a whole project afterwards.
Write the goal as one committed sentence.
The placeholder is a good model: e.g. Launch ToDidIt to 100 paying customers. Name the outcome, and add the deadline and the scale if you know them. Skip the adjectives, they buy you nothing.
Generate, and read the whole tree before touching anything.
It takes under a minute. Read top to bottom first. The instinct is to start fixing line one immediately, and it wastes edits, because the phase that is actually wrong is usually further down.
Decide whether the shape is right, not whether the words are right.
Wording you fix inline in seconds. Shape is the expensive thing. If the phases and their order look correct, keep the draft and edit. If the draft has misunderstood the goal, Start over with a better prompt rather than editing your way out of it.
What it costs
Generating and regenerating both cost 25 To-Dids. Refining costs 8. When the shape is close, refine. Regenerating to fix a phrase is paying 3 times over for a change you could have typed.
Every line is editable inline before you save: rename, reorder, re-prioritize, move dates, delete, add. Most drafts need ten or fifteen small edits and no AI at all, so start there and only reach for Didi when the change is structural.
Delete what you are not going to do.
Drafts run generous on purpose, because deleting a task is faster than thinking of one. Cut hard on the first pass. A plan you half-believe is worse than a shorter one you will actually work.
Fix names, priorities, dates and repeats inline.
Click any line to edit it. The repeat control sits beside the date and stays quiet until a task uses it. Add task inserts the thing the model could not have known about, which is usually something specific to your client, your stack, or your building.
Set or clear a repeat, in the editor or in chat.
Both work. Pick from the repeat control on the row, or say "change the audit task to repeat monthly" and Didi will do it. "Stop repeating X" clears one.
Use Refine with Didi for structural changes only.
"Split the launch phase into pre-launch and launch week" or "add a QA pass before every handoff" are refinements. "Rename task three" is not, and costs 8 To-Dids to do what a click does free.
Undo anything you regret.
Undo steps back through your edits. Nothing in the draft is precious and nothing is saved yet, so experiment freely.
Know when to Start over instead.
If you are more than about ten edits deep and still fighting the draft, the prompt was wrong. Start over costs another 25 To-Dids and will beat another twenty minutes of correction.
Saving is the only irreversible step, and even then it produces an ordinary project you can edit, archive, or delete like any other.
Read the first week one more time.
The near-term dates are the ones that will start pinging you. Everything past a fortnight out you will revise anyway, so do not spend care on it now.
Press Save & start.
The whole tree is written as a real project: tasks, subtasks, priorities, and dates. No import step, no staging area. Discard throws the draft away instead, and costs you nothing beyond the To-Dids already spent.
Open Today and confirm the near-term work looks sane.
Open TodayThis is the honest test of the draft. If Today is suddenly stacked with eleven items, the dates are too aggressive. Spread them now, while the project is still fresh in your head.
Assign the first phase if you are running it with other people.
A Blueprint saves unassigned. Handing out the first phase immediately is what turns a plan into work someone has agreed to do.
If you will run this shape again, save it as a template.
A Project Template costs no To-Dids to stamp out. Any project shape you repeat should be generated once and templated, not regenerated monthly.
Goal-shaped projects: launches, campaigns, events, moves, hiring sprints. Anything with a clear outcome and a sequence of steps. For open-ended exploration with no defined outcome, building it by hand is genuinely a better fit and will take you less time than fighting the draft.
It is also not a research tool. Blueprint drafts the structure of work you already understand. It does not know your client's contract terms, your team's actual capacity, or which of your vendors is unreliable, and a draft that looks confident about any of those is guessing.
- Good fit: a defined outcome, a conventional shape, a deadline you can name.
- Poor fit: ongoing operational work with no end state, like an inbox or a support queue.
- Poor fit: anything whose plan depends on facts only you hold.
- Poor fit: a project you have already run before. Use a template and keep the To-Dids.
A Blueprint is a first draft, not a verdict, and the useful part of the feature is what happens two weeks later. ToDidIt watches the project you saved and tells you where the draft was wrong, so the next one is better.
Run this as a short weekly review rather than a rolling anxiety. Ten minutes on a Friday is enough.
Check the Project Health Score.
Health reads completion rate, overdue load, and staleness together. A Blueprint project that drops fast in week two almost always had over-tight dates rather than a wrong plan.
Look at whatever the Stuck Sensor has flagged.
Stuck Sensor counts reschedules. A task you have moved three times is not a scheduling problem, it is a task that was drafted too big and wants breaking up.
Read the ToDid Log for what actually got done.
The Log is the record of real work against the plan you drafted. Where the two disagree, the Log is right.
Fix the plan, not the deadline.
Reshape the remaining phases by hand. This is ordinary project editing now and costs nothing. Only regenerate if the goal itself has changed.
When it finishes, template what worked.
The most valuable artifact of a Blueprint project is the corrected version of it. Save that as a template and the next run starts from something true rather than something drafted.
Tip
The pattern to watch for across projects is dates, not tasks. Blueprint drafts a reasonable sequence and an optimistic schedule. Most people only need to learn to add about 30 percent to the first estimate.
One generation costs 25 To-Dids. A refinement costs 8, and discarding a draft costs nothing beyond what you have already spent generating it. To-Dids are ToDidIt's AI credit: one monthly allowance per workspace, shared by everyone in it, covering every Didi feature rather than metering each one separately. See To-Dids for the full picture.
Squad includes 500 To-Dids a month, which is 20 Blueprints if you spent the entire allowance on nothing else. Business and Lifetime include 1,000, or 40. In practice nobody spends it all on Blueprint, because the cheap Didi features are the ones you use daily: a Smart Due Date is 1 To-Did and a description draft is 1.
The 14-day Squad trial includes three generations rather than the full monthly allowance. That is deliberate: three is enough to try Blueprint on projects you actually care about and judge it honestly, and not enough to plan a year of work on a free trial.
If you run out mid-month, a refill tops the workspace up immediately rather than making you wait for the reset, and Auto-Refill does it without you noticing. The allowance resets monthly and does not roll over.
What it costs
Every figure on this page is read from the same constants the app bills against, so a price change here cannot drift from a price change in the product. Today: 25 to generate, 8 to refine.
Four things account for nearly every Blueprint that disappoints someone.
- The draft is generic. The prompt was generic. Add the deadline, the scale, and the one constraint that makes your version of this project different.
- The dates are unrealistic. Expected. Blueprint drafts sequence well and duration optimistically. Fix the dates on save, and see Smart Due Dates for per-task help.
- It ran out of To-Dids mid-project. The budget is per workspace per month. Refills top it up immediately.
- The project went read-only. That is a downgrade to Free after the trial, not a bug. The project and its tasks are all still there. See what happens on downgrade.
The short answers. Each links to the full one.
- Can I edit a Blueprint after it is generated? Yes, and you have to: nothing saves until you press Save & start. Every line is editable inline, and after saving it is an ordinary project. More
- How many Blueprints do I get on the trial? Three generations across the 14 days, not the full 500 To-Dids. More
- Does it work for any kind of project? Goal-shaped work, yes. Ongoing operational work with no end state, no. More
- What happens to my Blueprint projects if I downgrade? They stay, and go read-only. Nothing is deleted, and upgrading restores editing. More
- Can I use Blueprint on the Free plan? No. It is Squad and up, plus three generations during the trial. More
Similar readings
Everything below is in the Happiness Center.
What is Project Blueprint?
A one-sentence goal in, a full project tree out.
How much does a Blueprint cost?
25 To-Dids per generation.
How many Blueprints do I get during the free trial?
Three.
Can I edit a Blueprint after it's generated?
Yes, every line, before anything is saved.
Does Blueprint work for any kind of project?
Best on goal-shaped work with a clear outcome.
What happens to my Blueprint projects if I downgrade to Free after my trial?
They become read-only. Everything else stays editable.
To-Dids: how AI usage is metered
A To-Did is the unit of AI cost. One Blueprint is 25, one Smart Due Date is 1. The meter in the top bar shows where you are.
Refills and Auto-Refill
Top up mid-month with a refill pack, nudge an admin to do it if you cannot, or switch on Auto-Refill and never see a pause.
Project Health Score
A chip on the project header reading On track, Falling behind, or Needs attention, with the reasons behind the verdict in a tooltip.
Stuck Sensor
Notices tasks you keep pushing and surfaces them as slipping, with a one-click snooze when something genuinely can wait. Free on every plan.
Project Templates
A saved snapshot of a whole project: every task, subtask, priority, and the relative spacing of the due dates. Launch a repeat initiative in one click.
The 14-day Squad trial
Two weeks of the full Squad tier, card required, including three Project Blueprint generations. Here is exactly what happens when it ends.