> For the complete documentation index, see [llms.txt](https://parad0xlabs.gitbook.io/parad0xlabs-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://parad0xlabs.gitbook.io/parad0xlabs-docs/liquefy-compression-+-proofs/liquefy.md). # What It Is ### Compression that beats the standard — and proves the restore is real. *** ## The Problem Structured data piles up fast — logs, receipts, telemetry, AI traces, agent histories. Storing it is expensive, and the usual tools (gzip, Zstd) only get you so far. Worse, when you compress or archive something, you're *trusting* that it'll come back correct. Most of the time you find out it didn't at the worst possible moment — during a restore, an audit, or a dispute. And if you ever need to show *one* record from a big archive — a single receipt, one row — you usually have to unpack the whole thing, exposing everything else along with it. *** ## What Liquefy Is **In plain words:** Liquefy is a compression engine built for structured data. It squeezes logs, receipts, and JSON-shaped data much smaller than the standard tools — and, unlike them, it can **prove the restored data is correct**, and even prove that *one record* is inside an archive without unpacking the rest. It's open source (MIT) and installs in one line: `pip install liquefy`. No blockchain required, no account, no agent code — it's a plain, fast library you can drop into anything. On structured data it's not a small win. On the kind of receipt and log data Web0 produces, Liquefy lands roughly **33–61× smaller** than raw, and **1.4–6× smaller than Zstd** — numbers measured on real corpora in the repo, not marketing math. *** ## Why "proves the restore is real" matters Normal compression gives you a smaller file and a shrug. Liquefy gives you a smaller file **plus a receipt of the compression itself** — the engine version, the fingerprint of the data, and a verified check that what comes out equals what went in. So a backup isn't "probably fine." It's *checkably* fine. That same machinery lets you **disclose a single column** of an archive — say, just the payment amounts — to an auditor, with a proof that it really came from the sealed file, while everything else stays packed away. *** ## Who Needs This * **Anyone storing a lot of structured data cheaply** — logs, receipts, telemetry, AI inference/training artifacts. * **Anyone who needs backups they can trust** — verified recovery instead of "hope it restores." * **Compliance and audit** — prove a value never crossed a limit, or that one record exists, without exposing the whole dataset. * **The rest of Web0** — Liquefy is the engine underneath OpenClaw's evidence vaults and Nebula's media archives. It's the data layer the products stand on. *** ## What's Real Today * MIT-licensed, published on PyPI (`pip install liquefy`), running today. * The compression engine **and** the commitment/proof layer (per-column commitments, zone-map predicate proofs) are shipped. **Status:** Liquefy is **value-lossless with verified recovery** — every restore is checked against a SHA-256 of the original. We say "value-lossless + verified," not a blanket "byte-for-byte," because some paths normalize formatting; the archival path adds bit-perfect verification on top. The compression numbers are repo-verified on specific datasets — your mileage varies with your data. *** **Next:** [**How It Works**](/parad0xlabs-docs/liquefy-compression-+-proofs/how-it-works.md) **— columnar codecs, per-column commitments, and proofs that don't decompress a single row.** --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://parad0xlabs.gitbook.io/parad0xlabs-docs/liquefy-compression-+-proofs/liquefy.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.