Lossless compression - byte-exact verified restore

Compress your data
without losing a single byte.

Pedulli is lossless compression with a byte-exact, SHA-256-verifiable restore and a +1 byte never-worse floor. It runs as a best-of-N racer that races xz, zstd, brotli and your data's SRD math and keeps the smallest verified output, so it is never larger than the best standard codec (worst case +1 byte). It wins outright on structured data and ties the best codec on already-optimal or random data.

input -> lossless compress -> .pdli container -> restore -> SHA-256 match

LESS WEIGHT - SAME BYTES - VERIFIED RESTORE

No lossy promises: every result must restore the original byte-for-byte, or it stays proof-only / diagnostic.

What Pedulli ships

Two products. Both verifiable.

Pedulli is the umbrella for two distinct products you can check yourself: the T15 Orchestrator for byte-exact lossless compression, and the SRD Engine Pedulli, a witness-gated exact verifier.

Product A Lossless compression

T15 Orchestrator

Byte-exact lossless compression with a SHA-256-verifiable restore and a +1 byte never-worse floor on its own output. Every input is read as exact bytes and turned into a restorable .pdli container; restore reconstructs the original and the SHA-256 must match.

It uses SRD math partly together with standard codecs (xz, zstd, brotli) and keeps the smallest verified output, so it is never larger than the best standard codec (worst case +1 byte). It wins outright on structured data and ties the best codec on already-optimal or random data. Not "superior to all", not "past the Shannon bound".

Product B Exact verifier

SRD Engine Pedulli

A witness-gated exact verifier. You supply a witness and the engine returns accept or a diagnostic - it checks what you supply against an exact relation across its verify.* domains: a real AES-GCM re-encryption check, a p*q == N bigint check, a SHA-256 restore roundtrip, supplied-clause SAT, and per-step proof checking. If the residual is zero it accepts; otherwise it returns a diagnostic.

Every capability is proof-gated: shipped only with a measured benchmark plus a published proof object; research tracks are labelled research, not delivered. It verifies witnesses you provide. It explicitly does not break or crack RSA, AES or ECC, and does not claim to solve every problem - it returns a supported answer with proof, abstains, or verifies a supplied witness across 1,428 domains. Internals stay sealed - proof without disclosure.

For teams that move a lot of data

Less weight. Same bytes. Verified integrity.

Pedulli helps companies, developers, AI and data teams, and storage-heavy infrastructure shrink storage, backups and transfers while keeping the original byte-for-byte restorable.

Media & streaming

Archives without re-encode

Compress heavy media files while preserving the original byte-for-byte. No lossy re-encode, no risky conversion.

AI & data teams

Verifiable datasets and logs

Shrink datasets, logs, backups and model files while keeping auditability, integrity and reproducibility.

SaaS & DevOps

Heavy logs and JSON

Lighten monitoring, exports, application events and structured data without complex manual configuration.

Enterprise storage

Cut storage, not data

Reduce archives, backups and transfers without deleting data, lowering quality, or rewriting your workflow.

API and CLI ready

Integrate byte-exact compression into your products without building a compression engine in-house.

No fake savings

Random, encrypted or already-compressed files stay proof-only when they cannot honestly shrink.

Provable roundtrip

Upload a file, compress, restore, and verify the restored output is identical to the original.

Try it - live - free

Drop a file - see how much smaller

Any file type or bitsequence. Byte-exact restore. Sub-1 MiB runs on our EU servers; larger files use the resumable upload API.

Drop any file or bitsequence here
MP4, JSON, logs, backups, datasets, source code, raw bytes, 010101 bitstreams
No storage - byte-exact restore - +1 byte max overhead
// ready - drop a file to compress
SESSION: 0 files compressed
Full trial + restore proof ->

Where Pedulli wins - and where it ties the best

Best-of-N racer, measured head-to-head

Bitcoin-anchored, SHA-256 byte-exact, reproducible. Because the racer includes xz, zstd and brotli and keeps the smallest verified output, it is never larger than the best standard codec (worst case +1 byte): it wins outright on structured classes and ties the best codec everywhere else.

Where it wins
  • Structured JSON / logs (about 2x tighter on structured benchmark samples)
  • Periodic and sparse binary structures
  • Specific MP4 master containers (sealed byte-exact benchmark)
Where it ties the best codec
  • Long English text: the racer selects xz, matching its size
  • Small generic text: the racer selects brotli, matching its size
  • Random / encrypted / already-compressed input (proof-only, worst case +1 byte)

Full protected pipeline

Any file to .pdli.
Then back to the exact original.

Self-sealed: proof without disclosure, engine internals remain closed, no absolute-security claims. Input bytes become a protected .pdli container, then restore reconstructs the original byte-for-byte.

01

Any bytes in

Bitsequence, PDF, MP4, ZIP, source, dataset, backup, log stream. The entry point is raw bytes.

02

Protected runtime

The protected runtime searches for the smallest reversible artifact. If structure exists it compresses; if the input is already closed it preserves exactness instead of inventing savings.

03

.pdli container out

The output is the public container: portable, API-ready, CLI-ready, designed for exact restore.

04

Inverse restore

Restore is not a best-effort decode. The inverse path reconstructs the original byte stream exactly.

05

Roundtrip proof

Restored bytes must match the original. No lossy mode, no approximation, no hidden mutation.

06

Diagnostic if not closed

If no smaller exact artifact exists, the result is proof-only / diagnostic - never a fake win.

COMPRESS

file.bin -> protected pipeline -> file.pdli

ROUNDTRIP

RESTORE

file.pdli -> inverse pipeline -> original file.bin

Pricing

Free for personal use. Pro for builders. Custom for enterprise.

Start free with an API key and token balance. Upgrade to Pro for the REST API, or top up tokens any time.

Free

Personal use

EUR 0 forever
Get free API key
  • No signup for the browser trial
  • Sub-1 MiB runs on our EU servers (deleted after the job)
  • Byte-exact restore + SHA-256 proof
  • +1 byte never-worse floor
  • Free API tier with token balance
Most popular

Pro

REST API for builders

EUR 49/month

or EUR 468/year (~EUR 39/mo)

Start Pro - EUR 49/mo
  • REST API + webhook + signed-URL callbacks
  • 10 GB/month included, per-GB after
  • Account and team API keys
  • Token top-ups: EUR 5 / 20 / 100 / 500
  • Pedulli CLI + REST API
  • Self-sealed: engine internals stay closed

Enterprise

Custom / on-prem

Custom
Book a call
  • On-prem / air-gapped / sovereign cloud
  • Source / OEM / strategic licensing
  • Domain-tuned deployment
  • Dedicated engineer

FAQ

Quick answers

How is Pedulli different from gzip / xz / brotli / zstd?
Pedulli runs as a best-of-N racer that includes those codecs and your data's SRD math, then keeps the smallest verified result. Because it includes the standard codecs, it is never larger than the best of them (worst case +1 byte): it wins outright on specific structured data classes, and on long English text or small generic text it selects xz or brotli and matches that size - a tie, never a loss. On random or incompressible data Pedulli returns a proof-only guard instead of pretending there are savings. It is not "superior to all" and it does not claim to pass any entropy bound.
Is my data byte-exact restored?
Yes. Lossless, byte-exact. SHA-256 roundtrip verified on every output. No lossy re-encoding, ever. The output never exceeds the input by more than one byte.
What if my file is already compressed or encrypted?
It takes the proof-only guard path: exact restore is preserved and no compression win is claimed. Not presented as savings.
What is the exact verifier?
A witness-gated check: you supply a witness and Pedulli verifies it against an exact relation (AES-GCM re-encryption, p*q==N, SHA-256 roundtrip, supplied-clause SAT). It verifies what you supply. It does not break or crack any cryptography and is not a general solver.
Do you store my files?
No long-term storage by default. Sub-1 MiB files run on our EU servers, deleted after the job. Server-side jobs use temporary processing space and delete job data after the response.
How do I verify your record claims?
Benchmarks are sealed via OpenTimestamps on the Bitcoin chain. The timestamp proves existence and measurement priority, not the engine internals. See /verify.

Pick your stage

From a first file to a signed contract.

Try it, get an API key, go Pro, or talk to the founder.

Fastest
Activation

Run the free trial

Drop a file in the browser, get a byte-exact .pdli, verify the SHA-256 roundtrip. No signup.

Drop a file ->
Signup

Get a free API key

Create an account, get an API key and token balance. Works without Stripe.

Start API trial ->
Paid

Go Pro

EUR 49/mo REST API with webhook callbacks and 10 GB/month included.

Start Pro ->
High-touch

Book a call

30-min intro for demos, enterprise / source licensing, and investor conversations.

Book 30-min call ->

Leave one line - the founder reads everything

Engine internals are never exposed in public replies.

Make your data smaller and auditable.

Free for personal use. Drop a file now and see the savings before you sign up for anything.

No credit card - zero storage on the browser trial - Bitcoin-anchored proof