Why cedarpad exists

Ideas in. Decisions out.

The bottleneck moved

For seventy years, building software was the expensive part. Every tool we work in — tickets, specs, roadmaps — exists to protect scarce builder time. Then building stopped being scarce. One engineer with an AI agent ships in an afternoon what used to take a sprint.

But watch a team of five, each vibe-coding with their own agent, and you see the new problem: they ship five afternoons of work in five directions. AI didn't just make building easy — it made misalignment cheap to produce at scale. The bottleneck moved upstream, to the oldest problem teams have: agreeing on what to build. Builders are everywhere now. Clear decisions are the scarce thing.

Chat is the wrong shape for deciding

The medium we got for working with AI is a chat thread: linear, single-player, gone by tomorrow. But almost everything worth deciding is structural — a system, a tradeoff space, three options and how they differ. Prose serializes structure; your head has to rebuild it. Whiteboards fixed structure but never finished anything: every session ends with someone photographing the board, and the real work restarting from zero somewhere else.

So we're building a board for deciding

Problems meet the team and its visual agents on one shared canvas. The agent answers in working things: research docs, diagrams, charts, custom apps, clickable prototypes. Everyone explores in parallel with their own agent; everything lands on one surface; the team converges by pointing at real things instead of arguing about adjectives.

Why this compounds

Every component your team builds becomes vocabulary — reusable words in how you and your agents communicate. A cedarpad workspace gets faster at aligning the longer it lives. A fresh chat window starts from zero every time.

Come build it with us

Cedarpad is invite-only while we build it, and we sit with every team we let in.