Tokens & savings

tokenolo wraps your coding agent behind a local proxy. Before each message is sent, it removes redundant filler from your own text and forwards everything else untouched. The effect on your bill is real but modest — and it never comes at the cost of your code leaving your machine.

You pay for every token you ship

Coding agents bill on the input the model receives. Some of that input is throwaway filler in your own prompts — "please", "could you", trailing pleasantries — that costs tokens without adding instruction.

tokenolo removes that filler before the request is sent. The tokens it strips are tokens you stop paying to transmit. It's a small reduction, not a dramatic one — which is the honest trade for keeping everything else (your files, tool output, and history) untouched and on your machine.

Note: we don't claim a fixed savings number. Filler-trimming is a modest, variable reduction — how much depends on how much filler your prompts carry. If you want a larger, server-side rewrite of a single prompt, that's the separate, opt-in tokenolo paraphrase command — not part of wrap.

Cache-safe by design

The edit is byte-stable, so wrapping never invalidates your provider's prompt cache. You keep your cache hits across turns — the trim is pure upside, never a cache miss.

How it ties to your plan

tokenolo pricing is quota-based. Every plan ships the same features with no gates; plans differ only by a daily wrap limit and a requests-per-minute ceiling.

Plan	Price	Wraps / day	Requests / min
Free	$0, no card	3	20
Pro	$19/mo	50	60
Team	$49/mo	500	120
Business	$99/mo	Unlimited	300

What this gets you

• Private by design — the proxy is local; your code never leaves your machine, and nothing is stored.
• Cache-safe — byte-stable edits keep your prompt cache valid across turns.
• A modest token trim — filler removed before send, on every run.

Next

• What gets optimized — exactly what tokenolo removes and what it forwards untouched.
• How it works — the path from your prompt to a trimmed input.