New — Meat Grinder (lvl 5) compression is now generally available

Pay for meaning, not formatting.

SHRNKR is a drop-in compression layer between you and your AI agent. Every prompt and response is rewritten into dramatically fewer tokens before it's billed — so the number on your invoice goes down. Immediately. Measurably.

Start shrinking — free $ npm install shrnkr

38%

avg. token reduction

2 lines

to integrate

companies trust us. 1 is us.

Live demo

See your savings in real time

Paste anything your agent might say. Watch it become cheaper.

Input — billed at full price

Output — what you pay for

Compression Vacuum Sealer (lvl 3)

How it works

Six proprietary compression passes

Each pass removes cost. None of them removes meaning. That's the whole trick.

Semantic Concatenation Engine™

Fuses adjacent words into a single compact token — same meaning, fewer characters on the invoice.

Adaptive Vowel Elision

Drops the vowels your model never needed to infer intent, scaled to how aggressively you want to save.

Sentence Flattening

Strips punctuation tokens that add cost without adding information.

Redundant Phrase Detection

Flags filler like "just wanted to check in" and "to be honest" and removes it before it ever reaches your model.

Symbol Substitution

Swaps common words for compact symbols your model already understands — "move" becomes "→", "and" becomes "&". Same instruction, fewer characters on the invoice.

JDSL™

Just Drop Some Letters. Once your response is squished into a single word, JDSL deletes every 4th character outright — no exceptions. Engages exclusively at Meat Grinder (lvl 5) tier.

Results

Optimized output, in production

Representative examples, prepared by our team for illustrative purposes.

Before

"Sure — here are three ways to reduce your AWS bill: right-size your EC2 instances, move cold data to S3 Glacier, and set up billing alerts."

After · Pressure Cooker (lvl 4)

Sr—hrarethrwaystordcurAWSbllrght-szurEC2instnces→colddattoS3Glcer&setupbllngalrts

↓ 42% chars · same three recommendations, none of the padding

Before

"I'd recommend scheduling the client call for Thursday afternoon, since Priya mentioned she's traveling Wednesday and won't be back until late morning."

After · Vacuum Sealer (lvl 3)

Idrecommnd schdlingthe clientcll forThursday aftrnonbc Pryamentioned shstravlng Wednsdy& wntbe backuntl latemrnng

↓ 25% chars · meeting still lands Thursday afternoon

Before

"The quarterly report shows revenue up 12%, driven primarily by growth in the enterprise segment, though churn ticked up slightly in SMB."

After · Meat Grinder (lvl 5)

thertryrptshsrenup2%dvnpmrlbygwthnthentprsgmnthgchrtckupsghtyinmb

↓ 52% chars · full report, roughly half the size

Pricing

We take 1% of what we save you

No subscription. No per-seat fee. SHRNKR's entire price is 1% of the token spend it removes from your bill. The harder it shrinks, the more you keep.

Your monthly token spend $4,000/mo

You keep (99% of savings) SHRNKR's fee (1% of savings)

Customers

What CTOs are typing

"Wecutourtknspndby41%inwkoneWeddntchangasnglprmpttmplt"

Dana R.

VP Eng at a Series B

"we approved the pilot expecting a rounding error in savings. it was not a rounding error."

Priyank M.

Founder

"half the responses read like a ransom note now. worth it."

Anonymous

On-call engineer

Install

Two lines. That's the whole integration.

Wrap any agent call. Savings begin on the next response — no config, no migration, no going back.

Try it first

$ npm install shrnkr

import { shrink } from 'shrnkr';

// wrap any call to your AI agent

const reply = await shrink(agent.respond(prompt), { level: 3 });

// reply: "sure,hersyourupdatd summry."