Stop paying for context your model doesn’t need.

Your prompts carry far more tokens than the model actually reads. Compresr drops the rest: up to 90% fewer tokens, so you cut cost and latency. At light compression it matches or beats full-context accuracy on public benchmarks.

Try the playground Talk about on-prem

See it on your own file in 60 seconds

Claude Code promptPaste into Claude Code

Use compresr to show me live cost savings on my own file.

1. pip install compresr
2. Introspect the SDK to discover the API, don't guess.
3. Ask me for COMPRESR_API_KEY (if I don't have one, send me to https://calendly.com/d/dvxq-cbv-5pz/intro-meeting-with-compresr to get access).
4. Ask me for (a) a path to a long document (PDF/.md/.txt) and (b) a question about it.
5. Compress the document with the question, then print a receipt: tokens in/out, ratio, GPT-5.2 cost full vs compressed at $1.75/M input tokens, % saved.
6. Ask GPT-5.2 the question against the compressed context and print the answer.

Don't skip the cost receipt; that's the point.

Works in Claude Code, Cursor, or any agent harness.Open full demo

How it works

We keep the signal and drop the noise.

Your raw text

Boeingreportedtotalrevenueof$77.8B202310-Kcommercialairplanes…+112,540 more

112,552 tokensBoeing 10-K, $0.263/query

compresr

compressing…delivered…

Keep the tokens that matter to your query.

Compression

226×

Compressed

revenue$77.8B2023

498 tokens$0.037/query, 86% cheaper

Tokens

112,552→498

226× fewer

Cost

$0.263→$0.037

86% cheaper

Latency

18s→13.7s

24% faster

What most teams are losing

Stop overpaying.

If you’re paying full price for your tokens, you’re leaving real money on the table.

~90%

Bill cut

vs. sending the full context

10×

Avg. compression

on long, sparse context

+3.7pp

Accuracy uplift

Compresr 2× vs full context, Pax arena

TodayWhat most teams are doing

Trimming / Truncation

Cuts off the tail: the answer was often in what you dropped.
Accuracy collapses on long docs.

Summarization

Lossy rewrite: nuance and exact wording are gone.
Costs extra LLM calls and latency for a worse context.

Question-agnostic compression

Compresses blindly: keeps irrelevant tokens, drops important ones.
Rarely gets past 5× without tanking accuracy.

With CompresrOne API call. Any scale.

Question-aware compression.

Feed us the query and the context. We return only the tokens that actually move the answer. You pay less, the LLM responds faster, and answers get sharper.

Per query: $0.263$0.037
Tokens in: 112,552498

Question-aware: we compress for the task.
At light ~2× compression, accuracy matches or beats full context.
SDK or on-prem. Your call.

Independent benchmark

FinanceBench.

	BaselineGPT-5.2	latte_v2 API+ GPT-5.2
Compression	None	~2x
Average Context	~106Ktokens	~56Ktokens
Accuracy	73%	77%
Savings	None	~47%cheaper

FinanceBench, 128 questions over SEC filings. At light ~2x compression accuracy holds; push to ~10x when cost matters more than peak accuracy

Two ways to deploy

Pick the one that fits your stack.

Hosted SDK

Drop-in SDK. One API key.

Install, add your key, compress any prompt or document before it hits your LLM. Volume pricing agreed up front, no surprise bills.

Volume pricing sized to your token spend
TypeScript & Python clients
Question-aware compression
Usage and spend per API key in the dashboard

On-Prem Deployment

Runs inside your VPC.

Your data never leaves your network. We deploy Compresr to your infrastructure, tune it for your workload, and support you directly.

Private deployment in your cloud or data center
Custom throughput & latency SLAs
Tailored to your business needs
Dedicated support

Enterprise, finance, healthcare, regulated workloads.