A wiki that gives AI context
I have set up LLM wikis for multiple clients: a maintained record of their projects, preferences, and decisions that their AI tools can return to.
The purpose
A new chat often starts with the same explanation of who you are, what you are building, and what was decided last time.
How it works
Existing work
Projects, preferences, and decisions
User-owned wiki
A small index with detail on demand
AI tools
Load relevant context in a new session
My part
I structured the wiki, connected it to each user's agent instructions, and guided them through keeping it useful.
In practice
Project context survives a chat session and can be reused by the user's configured AI tools.
A wiki needs upkeep. It gives the agent relevant context, not perfect recall or permission to act.
Want to know more?
Client and proprietary details stay private. Get in touch if you would like to discuss this project.
Ask about this project