staqd
All solutions
Knowledge & memory Flagship Coming soon

Knowledge Base

An agent-maintained second brain. Your notes, docs, and AI memory compiled into a living, cross-linked wiki.

The problem: Knowledge scattered across notes, project docs, and AI chat memory becomes unsearchable and forgotten.

Your knowledge, maintained by an agent

You accumulate knowledge everywhere — Apple Notes, dozens of project docs, AI chat memory, downloads. It piles up and rots. The Knowledge Base turns that pile into a single, synthesized, cross-linked source of truth that an AI agent keeps current and can answer questions against. A maintainable second brain instead of a dump of raw notes.

It's a finished product, not a prompt pack: a polished Next.js app plus the curated operating prompt that drives your agent to set it up, ingest your sources, and keep the wiki healthy.

How it works

  1. Harvest. A bundled connector pulls your sources — notes, project docs, AI memory files — into a raw/ folder. Read-only; your originals are never touched.
  2. Compile. You paste the curated setup prompt into Claude, Cursor, or Codex. The agent reads its operating manual, synthesizes raw/ into well-written wiki articles with frontmatter and [[backlinks]], and regenerates the index and knowledge graph.
  3. Browse. A dark-mode web UI gives you a dashboard, an article reader with "what links here", full-text search, a raw-source browser, and an interactive force-directed graph of how everything connects.
  4. Maintain. Ongoing compile, lint, and Q&A are all agent-driven through ready-to-run prompts — not manual gardening.

What you get

  • The complete Next.js 16 template (App Router, React 19, Tailwind v4) — dashboard, article viewer, search, raw browser, and graph view.
  • The curated operating manual (CLAUDE.md) — the specialized prompt that teaches your agent how to compile, format, link, lint, and answer.
  • Ready-to-run prompt scripts for harvest / compile / lint / Q&A.
  • Source connectors to get your material in with one command.
  • Lifetime updates to the solution.

Why it's different

Files are the database — everything is plain markdown in Git, so there's no database, no SaaS to depend on, and nothing to lock you in. The agent does the maintenance, and it synthesizes rather than copies, so the result reads like a well-written reference wiki, not a transcript of your notes.

Heads up: the Knowledge Base is in final productization. Join the waitlist and you'll be first in line at launch — Studio members get it included automatically.

What you get

  • Files are the database — plain markdown in Git, no DB, no lock-in
  • The agent compiles raw notes into a synthesized, cross-linked wiki
  • Backlinks, related links, and an interactive knowledge graph
  • Full-text search + a polished dark-mode reading UI out of the box
  • Ships with the curated operating prompt — point it at your own sources
  • Private by default, with explicit secret-redaction rules

Setup in minutes

  1. 1Drop the template into a folder and install dependencies
  2. 2Run the harvest script to pull your notes, docs, and AI memory into raw/
  3. 3Paste the setup prompt into your agent and let it compile your wiki
  4. 4Run the dev server to browse, search, and explore the graph

The curated setup prompt

You are the maintainer of this Knowledge Base. Read CLAUDE.md for your full
operating manual. Compile the raw material in raw/ into the wiki/: synthesize
(don't copy) each source into clear, cross-linked markdown articles with
frontmatter and [[backlinks]], then regenerate the index and graph. Never
store secrets. Begin by reading wiki/_index.md to see what already exists…
Preview. The full setup prompt is delivered with your purchase.

More solutions

Marketing siteLive

Launch Kit

A free starter solution — the curated prompts and section blueprints to scaffold a polished landing page in your AI tool.

Next.js 16 (App Router)Tailwind CSS v4Any staqd design skill
Free
in Studio
Customer supportComing soon

Support Copilot

A self-hosted support agent that answers from your own docs, with human handoff and a clean inbox UI.

Next.js 16Postgres + pgvectorYour LLM provider
$149one-time
in Studio
Marketing & SEOComing soon

Content Engine

An agent-run content pipeline — research, draft, edit, and publish SEO articles on a schedule, from your own brief.

Next.js 16PostgresCron / scheduler
$199one-time
in Studio