SILKLEARN Is Opening Early Access

We built SILKLEARN to solve a problem that kept surfacing in technical work: dense source material — internal docs, runbooks, codebases, technical specs, domain knowledge — with no reliable way to see what to read before what. The knowledge exists. The path through it doesn’t.

SILKLEARN changes that. You feed it source material, and it produces a dependency-ordered reading path — sequences where each concept builds on what came before, not a flat list of topics someone arranged by gut feel.

Today we’re opening early access.

Why Now

We’ve been running SILKLEARN internally and with a small set of design partners for the past several months. The core engine — decomposition, dependency mapping, and path generation — is working. People using it report spending less time figuring out what to read in what order, and more time actually in the material, on paths that reflect the real structure of the knowledge base rather than the date each document was written.

We’re not opening access because the product is finished. We’re opening access because the next round of improvements needs to come from real people using it on real source material — across domains and source types we haven’t touched yet. The signal that matters only comes from production use.

What You Get

Early access is not a demo or a sandbox. You get the working product.

That means: upload your source material — PDFs, documentation, runbooks, specs — and SILKLEARN maps the dependency structure across them. You get a reading path, not a reading list. You get flagged contradictions between sources, not a pile of documents you have to cross-reference manually. You get outputs organized for how you’re actually going to use the material, not how whoever wrote it assumed you would.

Early access also means direct access to the team. We read every piece of feedback. The roadmap for the next six months is going to be shaped by what people find useful and what they run into first — and we want that signal from people who are genuinely trying to use this on real material, not people looking for a clean demo environment.

Who This Is For

Early access is for people with a specific problem: source material that exists but doesn’t have a navigable structure through it. Engineers trying to onboard to a system that was never fully documented. Technical leads trying to turn internal knowledge into something they can hand off without a two-hour explanation. Anyone trying to learn a domain that has no canonical reading order — just a pile of primary sources, recorded calls, internal specs, and expert intuition that lives in someone else’s head.

If that describes you, the waitlist is open.

We’re limiting the first cohort deliberately — not to create artificial scarcity, but because the quality of feedback degrades at scale, and we’d rather have a hundred people who are actually trying to use this than ten thousand who signed up to see what happens. The access we’re granting now is the access that shapes what this becomes.