One Bootstrap per Conversation
PCL-Lite
The World's First Programming Language for AI Persona Management
Make AI behavior programmable, portable, and predictable.
We recommend using one bootstrap per conversation. This keeps the model focused on one domain (Standard, Image, Video, Audio, or Slides), so personas and constraints stay coherent and results are easier to optimize.
- One conversation = one set of personas and one context (e.g. image or video). No mixing of registries, no conflicting constraints.
- The model can follow one clear set of instructions.
- A single bootstrap uses context and tokens efficiently.
When you need another domain, start a new conversation and paste the bootstrap that matches:
- Need a video script → new chat, paste bootstrap-video.
- Need an image prompt or thumbnail → new chat, paste bootstrap-image.
- Need architecture or dev → new chat, paste bootstrap Standard.
- Guides — How to use each bootstrap.
- Getting Started — Quick start.
- Home