A token saver for your coding agent. Same output, 60-90% fewer tokens spent on verbose commands.
Measured, recoverable & never lossy where it matters.
How it works
Dev commands return a compressed, high-signal version of their output.
Not just your terminal. A localhost proxy sits inside your app’s own calls to Anthropic and OpenAI.
Reports tokens saved per command and session. Measured, and labeled honestly.
See it in action
Click “Run demo” to see a real git diff filtered before it reaches the model.
Same information, filtered before it reaches the model. Run julius savings for the full report; verbose commands typically compress 60 to 90%.
Built for application-level savings
A localhost proxy sits inside your app’s own calls to Anthropic and OpenAI, catching what CLI-only tools miss.
Point ANTHROPIC_BASE_URL at Julius. No SDK changes, no redeploys, works with any existing app.
No account, no vendor lock-in. Inspect the filters and run it entirely on your machine.
Need this enforced across the whole fleet, with audit and policy on top?
hoop is the same wire-level interception, running centrally, with access policy, data masking, and full audit logs across every agent, every app, and every LLM call your org makes.