C9AI Demo Lineup
High‑impact, 10–12 minute flow
Setup (2–3 min)
Start local stack
c9ai stack -m /path/to/model.gguf -p 8080
Opens Agent API at http://127.0.0.1:8787 and checks llama.cpp health.
Set provider (optional)
# macOS/Linux
export LOCAL_PROVIDER=llamacpp
# Windows (PowerShell)
$env:LOCAL_PROVIDER = 'llamacpp'
1) Natural → Actions (3 min)
Interactive CLI
c9ai
- create a file notes.txt with 'hello team', then show it
One‑shot
c9ai agent "list files changed today in this repo"
Shows planned steps, then shell/fs tools.
2) Local AI + Tools (3 min)
c9ai agent -p llamacpp "show my node version and top 5 dirs"
c9ai agent -t 0.8 "remove temp_test_module.js if present"
Privacy: runs fully local with llama.cpp.
3) Executive Calculator (2 min)
c9ai agent " @calc x=1. y=2. return x+y"
c9ai agent " @calc npv(0.08, [-100000, 30000, 30000, 30000, 30000])"
Deterministic, structured results (no hallucinations).
4) Latest Headlines (2–3 min)
c9ai agent -a "web.search,rss.read" "fetch top headlines"
UI shows big titles, date + publisher, full‑width thumbnails, pill‑style Read more.
5) Quick File Ops (2 min)
c9ai agent 'jit file operation=count-lines input=README.md'
c9ai agent 'jit file operation=extract-emails input=README.md'
c9ai agent 'jit file operation=word-frequency input=README.md'
6) Start Stack via Shortcut (1 min)
c9ai-stack
Equivalent to c9ai stack; ensures llama.cpp + Agent API are up.
Optional Extras
- GitHub issues:
c9ai todo fetch-github -r owner/repo -l bug
- Hybrid provider:
c9ai agent -p claude-hybrid "summarize README into bullets"
- CSV analyzer:
c9ai agent 'jit analyze file=data.csv'
Tips
- Have a GGUF model ready (Phi‑3 mini) and verify llama.cpp health in advance.
- Keep terminal + UI visible; switch providers on the fly when relevant.
- If network is flaky, stick to
-p llamacpp and local tasks.