What Is Knolo? Open-Source Verifiable Knowledge Images Explained

Knolo is an open-source, local-first Knowledge Image runtime and file format. You pack documents into one portable artifact, pin it by digest, mount it in your own process, and retrieve evidence without standing up a separate index service. The reference runtime is @knolo/core on npm (currently 6.2.0), released under the Apache-2.0 license. This post is a plain map of what Knolo is, what a Knowledge Image contains, and where to start on Knolo on knolo.dev.

Knolo in one sentence

Knolo turns source material into a bounded, content-addressed Knowledge Image you can copy, mount, query, and verify locally. The same image should produce the same retrieval behavior wherever @knolo/core mounts it, because identity lives in digests and roots rather than in a floating remote index.

That framing shows up across the public pages: Knowledge Images are portable evidence containers, not another ranked list behind an opaque API. You own the host application—networking, models, and product logic stay outside @knolo/core.

What is a Knolo Knowledge Image?

A Knowledge Image is a packed artifact of knowledge state: documents, structure, and metadata bound so another machine can inspect and check the same bytes later. V5 defines the current container contract in source and docs: dual superblocks with generation-based recovery, immutable object and event segments, a commit record with domain-separated roots, per-section bounds and digests, and fail-closed validation when required sections or digests do not match. The Knolo V5 concepts page is the developer-facing overview.

Because the image is content-addressed, pinning a digest means you refer to one immutable artifact—not “whatever the wiki said last Tuesday.” That is the core product promise: knowledge you can pin, mount, and compare.

Build, mount, and query with Knolo locally

The everyday loop with @knolo/core is build → mount → query. Lexical retrieval is the default path: deterministic, reproducible, and independent of a vector index. A minimal sketch looks like the examples in the Knolo direct retrieval docs:

import { buildPack, mountPack, query } from '@knolo/core';

const docs = [
  { id: 'policy.refunds', namespace: 'support', heading: 'Refunds',
    text: 'Refund requests must be submitted within 30 days.' },
];
const bytes = await buildPack(docs, { semantic: { enabled: false } });
const pack = await mountPack({ src: bytes });
const hits = query(pack, 'How long do I have to request a refund?', {
  topK: 5, namespace: 'support',
});

Install with npm install @knolo/core and follow the Knolo installation guide. The same package works in Node, the browser, and mobile hosts that can load the runtime. Optional semantic reranking exists for paraphrase cases, but it sits on top of lexical grounding—it does not replace it, and it does not require an external vector database. Yesterday’s deeper comparison stays on the site as Knolo vs a vector database; the short version for this overview is simply: Knolo is not a vector database and does not need one.

Related packages on npm today include @knolo/mcp 0.2.0 and @knolo/code 0.1.1. Cite versions from the registries; do not assume unpublished names are launched.

What Knolo verification means

Knolo verification answers a narrow question: do these bytes match the roots they declare? TypeScript and Rust share fixtures for deterministic encoding and SHA-256 digests. A passing check means the artifact is intact relative to its declared roots. It does not mean a source document is correct, a claim is true, or a model answer is right—the docs state that boundary on purpose.

Fail-closed validation is part of the V5 design: missing sections, overlaps, and digest mismatches stop the mount rather than silently trusting broken bytes. That is how a Knowledge Image stays useful under review: you can point at a digest and a verification result, not only at a chat transcript.

Where Knolo Hub fits

Public Knowledge Images are listed on Knolo Hub (“Knowledge you can pin.”). You discover a pack, inspect its card, pin a digest, and mount the bytes with @knolo/core on your machine. Hub is a registry for pin and mount—not the place queries run. The developer Hub docs live at Knolo Hub docs.

Start with Knolo today

1. Read the product overview on Knolo on knolo.dev.
2. Install @knolo/core 6.2.0 from npm and walk through installation.
3. Build a small pack, run query(), then verify the artifact with the verification guidance.
4. For open specifications, licensing, and the public record, use Knolo on knolo.org.
5. Source, issues, and releases: the Knolo core repository on GitHub.

Knolo is open-source infrastructure for portable knowledge: Apache-2.0, local-first, pin-by-digest, verify before you trust the bytes. That is the shape of the project, and the pages above are the live places to confirm it.