Guides
Practical guides
Short, task-oriented paths through the product. Use these when you want to do something specific, not read the whole manual.
Set up Cursor
Point Cursor at the binary, choose agent mode, and call tools in the right order.
Run on Jetson Orin
Use the Jetson package and connect Cursor over a trusted LAN.
Keep it local
Understand SQLite storage, loopback defaults, and when remote bind is appropriate.
Ask for implementation context
Shape tasks and budgets so agents get cited packages instead of sprawling dumps.
First ingest (any platform)
After you launch Librarian or have a database path ready, ingest something you own:
# Markdown / docs tree
./ingestcli -db ./implcache.db -mode markdown -root my-docs -path /path/to/docs
# Application or SDK source tree
./ingestcli -db ./implcache.db -mode project -root my-app -path /path/to/srcOn Windows packages, use ingestcli.exe and PowerShell paths instead. Prefer clear root names—agents will see them when a query needs a corpus choice.
Suggested learning path
- What is ImplCache
- Try it in 60 seconds on the product page
- Cursor MCP server
- Implementation context
- ImplCache vs RAG if you are evaluating architecture choices
Need the exhaustive operator manuals? Start at the docs hub or the Users Manual.
Ready to try it? Packages ship for Windows, Linux, and Jetson.