Skip to main content

​
The Evolution

Olly started as a terminal-based coding assistant — a local alternative to cloud AI that keeps your code on your machine. But developers don’t just live in terminals. Today, we’re announcing Olly’s next chapter: a web interface with real-time streaming.

​
Whatlaunched

​
REST API + WebSocket Server

The same agent that powers Olly’s terminal interface now exposes:
  • REST endpoints for health, state, and chat
  • WebSocket for real-time event streaming
  • Async processing — send a message, get instant confirmation, watch the response stream live

​
The Local-First Promise

Olly was built on a simple premise: your code stays on your machine. No cloud APIs, no data leaving your environment. The web interface extends this promise:
  • Host it locally or on your own infrastructure
  • No account required
  • No usage limits
  • Complete data sovereignty

​
Why This Matters

​
Privacy-First Developers

Every week, there’s another headline about code leaking into AI training data. Enterprises are increasingly restricting cloud AI tools. Olly answers both concerns:
  • Code never leaves your machine
  • Model runs locally (llama-cpp-python)
  • Self-host anywhere

​
Seamless Workflow

Whether you prefer the speed of CLI or the comfort of a browser, Olly adapts to you. Same agent, same tools, same policy engine — two interfaces.

​
What’s Coming Next

Phase 2 brings the Web Chat UI:
  • Next.js frontend
  • Real-time token streaming
  • Message history
  • Multiple concurrent sessions

​
Try It Today

cd olly
make serve
# Open http://localhost:8090
Or stick with the CLI — Olly meets you where you work.
Related: Technical implementation details | Business context