Every team,
on one canvas each.
A product team of AI employees per workspace, a canvas the whole company can read, and a build that lands in your repo.
- A workspace per product
- Its own canvas and AI team
- Everyone reads it
- Stakeholders comment, no seat
- Your repo
- The code and Postgres are yours
- Your model, or ours
- Set per workspace
Four teams, four readings of one document.
Every handoff is a translation. Every translation is a sprint.
- ProductWrites the PRD
- WaitDesign pictures it
- DesignPrototypes, in Figma
- WaitEngineering interprets it
- EngineeringBuilds its reading
- WaitProduct finds out in QA
“The PRD said ‘let users cancel a pending swap’. Three teams built three versions of what that meant.”
Product writes the PRD, design pictures it, engineering interprets it, and QA finds the gap. Here the spec is drawn as boards the moment it exists, everyone reads the same canvas, and the build ships exactly those boards. Your seats direct. The AI employees do the craft.
One canvas per product. Everyone on it.
Shared, edited live, and read by the people who never open Figma.
Notes land on the boards. Nadia turns them into changes.
A note goes on the exact board. It reaches the AI employee who fixes it, and the question it raised waits for whoever owns it.
Your design system and your voice, set once, on every canvas.
Type, palette, the words you use and the ones you do not. Every new workspace starts from them, so the tenth product looks and reads like the first.
Your model account, or ours. Per workspace.
Where the prompts go is a setting, and the code lands in your repo either way.
On your own model
your Claude or Codexthe team
On the account you already have. Your prompts go to your own provider. Nothing metered by us.
Your accountInside the workspace
the dashboardthe team, on our API
No keys to manage. The team signs in and works. Priority builds, seats per person.
Talk to usA repo in your org. A live URL.
Next.js and Postgres, committed to your GitHub as it goes. Priority builds.
Enterprise: common questions.
What does HuntingHydra for enterprise include?
A workspace per product, shared canvases edited live, one board for everyone on the team, priority builds, and seats billed per person. Every canvas has the whole team of AI employees on it.
Where does the code live?
In your repo, from the first build. Postgres is yours. Nothing is locked in the product.
Where do our prompts go?
On your own model account, to your own provider over MCP, with nothing metered by us. Or run the AI employees in the workspace on our API. Your choice, per workspace.
Can stakeholders read the canvas without a seat?
Yes. Anyone with the link can look and leave notes on the boards. Seats are for the people who direct the work.
Can our design system and voice be set once?
Yes. Set the type, the palette, the voice and the rules on the workspace, and every new canvas starts from them.
What does it cost?
The $20 plan is self serve, three days free. Team and enterprise seats are billed per person. Talk to us for the number that fits your team.
Every team, on one canvas each.
Try it on one product, three days free. Then talk to us about the rest.