Knowladex compiles your notes, transcripts, and scratch files into a real linked wiki — then plugs it straight into Claude, Cursor, and any tool that speaks MCP. Your knowledge, in the loop with your agents.
Knowladex is in private beta. Sign-ups are invite-only for now — apply below and we’ll let you in as we onboard batches.
Get raw files in, let an LLM compile them into a wiki, then read it — or let your AI tools read it for you, live, over MCP.
Drag & drop a folder, run the CLI, or have your agent ingest files for you over MCP. Notes, transcripts, specs, PDFs — anything text. No tagging, no folder ceremony, no schema up-front.
your scratch folder
notes/
├── q1-retro.md
├── customer-call-march.txt
├── product-spec-v2.md
├── strategy-thoughts.md
└── + 144 more
Tell Claude (or whichever AI you use) to make sense of it. It reads everything, finds the patterns, and writes a real wiki — connected articles, clean cross-references, sources cited.
you › organize my knowledge base claude › reading 147 documents… ✓ wrote customer-feedback ✓ wrote q1-decisions ✓ wrote product-roadmap ✓ 12 articles, 38 cross-references
Knowladex is an MCP server, so Claude, Cursor, ChatGPT Desktop, and the rest can search, read, and write to your wiki directly. Ask plain-language questions; get answers with sources, every time.
you › claude, what did we decide about pricing in Q1? claude › in March you settled on three tiers with a one-time setup fee. sources: [q1-decisions] [pricing-thread]
Not a PDF dump, not a chatbot. A real knowledge product with editorial typography, a working MCP server, and the ergonomics your team will actually use.
Editorial typography, cross-linked articles with backlinks, hover previews, tags, and a full A–Z index. Every article is a markdown file with frontmatter — readable as-is, no proprietary format. Looks like something a publisher would ship, not something a model spat out.
Drop in raw notes, transcripts, half-finished specs. Ask your AI tool to compile. It reads everything, decides what articles to create, and writes them back as a real interconnected wiki — with wikilinks, sources, and structure. You bring the chaos; the wiki appears.
Two ways in. Point Claude Code, Cursor, ChatGPT Desktop, or any MCP-speaking agent at your workspace and they can ingest, search, read, write, and compile in plain language. Or just drag a folder into the dashboard. Same backend either way.
Native PDF ingest with two-pass figure extraction — embedded bitmaps and vector-drawn diagrams that other tools miss. Text and figures land inside the compiled articles automatically, on the right pages. Up to 22 MB per file, no OCR shuffle.
Flip a project to public and send a clean read-only link
— knowladex.com/share/<org>/<project>.
Search and a tag cloud included, no signup needed for
readers, raw docs stay private. Toggle off any time.
Download your entire compiled wiki as a zip — every
article as a markdown file with frontmatter, an
.obsidian/app.json so wikilinks resolve, and
an index. Open it in Obsidian, edit it locally, take it
anywhere. Your data is never trapped.
Each client gets their own private workspace, their own team logins, and (on bigger plans) their own branded URL. You sit above all of them. Onboarding a new client takes minutes.
Stop pretending anyone has time to update Notion. Drop your meeting notes and decisions in — your AI keeps the wiki current, so the next person to ask a question gets a real answer.
One folder. One pile of half-finished thoughts. Wake up to a wiki that organizes itself, every time you add to it. It compounds, like compound interest for your brain.
Three tiers. Cancel any time, no commitment, no setup fee, no per-seat add-ons. Pricing will be announced when the beta opens — apply below and we’ll lock in your tier when seats open.
Every tier includes the full MCP server, the dashboard, the wiki view, public share URLs, Obsidian export, and PDF ingest with figure extraction. The difference between tiers is how many workspaces and documents you get and how fast you can push the system — not which features you can touch.
Knowladex is in closed beta while we onboard the first batches of operators, agencies, and AI-native teams. Leave your email below and we’ll let you in as seats open up. No credit card, no pricing surprises, no “marketing sequence” nonsense — just a confirmation email and a follow-up when it’s your turn.
We set up your workspace, move your first batch of docs in, and hand you the keys. You bring the chaos. Private beta for now — apply below and we’ll let you in as we onboard batches.
Apply for the betaAlready have an invite? Sign in