Solution designs your agents can read
Describe how you want to build it in markdown. Link each design to the features and ADRs it relates to, and let AI agents read the intended design over MCP before they write code.
Everything you need to author, connect and share architecture
Markdown-native authoring
Write solution designs in markdown with a title, summary and structured body. Code blocks and sections describe your architecture clearly.
Clear status at a glance
Move each design through draft, review, approved and superseded so the team always knows which architecture is current.
Linked to features and ADRs
Connect designs to the features they enable and the architecture decisions that shaped them. Full traceability across your spec graph.
Readable by AI agents
Agents fetch the markdown over MCP with get_solution_design and list_solution_designs, so they read the intended design before building.
AI drafts from your documents
Import text or PDF documents and AI drafts entities from them. Each draft arrives as a suggestion you accept or reject.
You stay in control
AI-drafted entities are pending until you review them. Accept what is right, reject the rest, and only approved content becomes part of the spec.
How it works
Write your design
Author a solution design in markdown describing your architecture: endpoints, data models, auth flows and infrastructure.
Link it to the spec graph
Connect the design to the features it enables and the ADRs that shaped it, so every choice stays traceable.
Set its status
Mark it draft, review, approved or superseded so everyone knows which architecture is the current one.
Share with the team
Stakeholders read the design in the web portal. AI agents read it via MCP before writing code.