Plans don't have to be flat.

We built Topolog because every productivity tool we used asked the same wrong question. What's on your list today? But the tasks on your list aren't independent. They depend on each other. They take order. They have a critical path.

Engineers and project managers have been planning around dependency graphs since the 1950s. PERT-CPM, Gantt charts, sprint planning; all built on the insight that order matters. None of those tools are built for personal goals.

Topolog is. It's an AI-powered planner that takes your goal, models its dependencies, and schedules your real days around them. The AI generates structure. The scheduler is deterministic code. Your job is to mark things done.

We don't ship the naive version of anything

Every layer of Topolog exists because the obvious version of it doesn't work. We tried the obvious version, watched it fail, and built what the problem actually needed. That pattern is the whole product.

• The obvious planner is a flat list. But your tasks aren't independent; they depend on each other and take an order. So we model every plan as a dependency graph.
• The obvious forecast is a single date. One date is a guess that's wrong the moment you write it. So we run a Monte Carlo over the whole plan and show the full completion spectrum: your odds of landing on time, partially, or missing.
• The obvious AI hands the goal to the biggest model. We tried that first, and the largest model available couldn't produce a single valid plan. So we built constrained hierarchical authoring: decompose top-down, validate every step, repair or drop anything that fails. The first real run produced a 37-milestone plan with zero validation errors.
• The obvious AI product lets the model do the math. That is how forecasts hallucinate. So the AI never touches the numbers; the scheduler and the Monte Carlo are deterministic code the model can't reach.

Who's behind it

Rohith B.V.

Senior software engineer. Previously Meta, HubSpot, HackerRank, Instamojo. IIIT Hyderabad CS. Left FAANG to build the planner I actually wanted to use.

What we believe

• → The user is the architect. AI is the draftsman.
• → The scheduler is a deterministic algorithm, not a chat prompt.
• → Slips happen. The plan should absorb them silently.
• → Your data is yours. Export to JSON anytime.

How to reach us

Drop us a note at hello@topolog.co.uk. We read everything.