Part of the Coding-Autopilot-System ecosystem: gsd-orchestrator | autogen
Promptimprover is an MCP server middleware that intercepts and refines every AI prompt before code generation — applying project context, coding standards, and compounding memory.
- RAG neural snippets — FlexSearch-based retrieval over the local codebase; injects relevant code examples into every prompt
- Compounding memory — SQLite-backed pattern store accumulates project-specific rules and standards learned over time
- Auto-heal middleware — background file watcher triggers commit ingestion and lesson extraction; keeps context current without manual intervention
- Context-aware project scouting — NodeDetector, PythonDetector, and ArchitecturalScout identify tech stack and patterns at startup
flowchart LR
CLI["AI CLI\n(Claude / Cursor)"] -->|"stdio"| PI["Promptimprover\n(prompt-refiner)"]
subgraph internal["Promptimprover Engine"]
RAG["RAG Snippets\n(FlexSearch)"]
Memory["SQLite Memory\n(LocalBrain)"]
AutoHeal["Auto-Heal\n(BackgroundService)"]
end
PI --> RAG
PI --> Memory
PI --> AutoHeal
internal --> Out["Augmented Prompt"]
git clone https://github.com/Coding-Autopilot-System/Promptimprover.git
cd Promptimprover
.\build_and_install.ps1Add prompt-refiner to your MCP client configuration. See the Setup Guide for full configuration instructions.
MIT — see LICENSE