Trurlic Documentation
Trurlic captures your architectural decisions in a typed graph inside your repo and serves them to every AI coding agent as hard constraints — before a single line is generated.
Think before you build. Then let the machine build what you thought. Named after Trurl from Stanisław Lem’s The Cyberiad — the constructor who reasons deeply about what he creates before creating it.
Trurlic is an MCP server. Your coding agent talks to it directly: it pulls architectural context before implementing, runs design conversations when it hits new territory, and records decisions as it goes. The trurlic CLI is for you — initialize, inspect, and maintain the graph, and open the visual map. The CLI never calls an LLM; all reasoning runs on your agent’s own model, over MCP.
Start here
Section titled “Start here”| You want to… | Go to |
|---|---|
| Understand what Trurlic is | Overview |
| Install and record a first decision | Getting Started |
| Learn the model | Core Concepts |
| Connect your coding agent | MCP Server · Connect Your IDE |
| Map an existing codebase | Bootstrap |
Documentation map
Section titled “Documentation map”Guides — MCP Server · Decision Graph · Workflow Engine · Design Conversations · Bootstrap · Patterns · Interactive Map
Architecture — System · Integrity Model
Reference — CLI · MCP Tools · Troubleshooting
Integrations — Connect Your IDE · Claude Code · Custom Agent