This estimator runs the same decision rules as the live proxy: the 12,000-token floor, the keep-the-last-4-messages-verbatim rule, and verbatim preservation of code and structured data. If your traffic is a bad fit, it says so.
Monthly math: we assume a conversation shaped like your paste happens [conversations/day] times daily, and that each conversation re-sends its growing history once per turn. We approximate that sum as (turns ÷ 2) × total tokens per conversation, which slightly overstates early turns and understates late ones. Subscription fees ($29/mo Pro, $199/mo Agency) are not included in the numbers above.
Estimates use the same thresholds as the live proxy but real savings depend on your content; the free tier's $5 credit is the real test.
Those requests pass through untouched - model your savings as zero and keep your cache. We never break a cache to force savings. Full comparison here.
No black box. These are the same rules the live proxy applies, minus the summarization model itself.
~1,500 tokens each.12,000 tokens, the proxy forwards it untouched. You pay nothing extra and save nothing.4 messages are always forwarded verbatim. The model needs full fidelity on what was just said.{ or [), base64 runs, images, and structured blocks (tool calls, tool results) in older history are preserved verbatim, never compressed.0.4 target ratio: forwarded = protected + 0.4 × compressible + last 4 messages.The free tier's $5 credit against your live traffic beats any simulation.