Desk with laptop, sticky notes, and sunlight — Photo by Richard Stachmann on Unsplash

Photo by Richard Stachmann on Unsplash

You've probably had this experience. You read a great article, highlight the key ideas, paste them somewhere — and three months later you can't find them. You know you've read something about this topic before. You know you have notes on it. They're somewhere.

The problem isn't that you're not taking enough notes. It's that your knowledge isn't connected, searchable, or useful at the moment you actually need it. A Personal Knowledge Base (PKB) fixes that. And in 2026, combining Obsidian, Notion, and Claude turns it from a storage system into something that actively helps you think.

This guide covers the setup, the workflow, and the honest pricing at each tier — so you can start without overbuying tools you don't need yet.

What Is a Personal Knowledge Base?

A PKB is a system for capturing, organizing, and retrieving everything you learn — articles, ideas, research, meeting notes, shower thoughts, half-formed hypotheses. The goal isn't to store more. It's to make what you've already learned useful: findable, connectable, and ready to inform your next decision.

The concept has been around for decades — Niklas Luhmann built a 90,000-card Zettelkasten that generated 58 books. But the tools available in 2026 make a genuinely useful PKB accessible to anyone willing to spend an afternoon setting one up and about 10 minutes a day maintaining it.

The goal isn't to store more. It's to make what you've already learned useful when you need it.

The Three Tools and What Each Does

These three tools serve distinct roles. Using all three for the same thing is the most common mistake people make — so before we talk about workflow, it's worth being precise about what each tool is actually for.

Obsidian — Your Private, Local Thinking Layer

Obsidian is a note-taking app that stores everything as plain Markdown files on your own device. No cloud required, no subscription, no company that could shut down and take your notes with it. Your files are just files — you can read them with any text editor and sync them with any service you trust.

Its killer features are the graph view (a visual map of how your notes link to each other) and the bi-directional linking syntax ([[double brackets]]), which lets you build a web of connected ideas rather than isolated documents.

Obsidian Pricing:

  • Personal use: Free forever. No limits, no nag screen, no watermark.
  • Commercial license: $50/year (~$4/month) — required only if you're using it inside a for-profit company with more than $1M in annual revenue. Most people reading this: irrelevant.
  • Obsidian Sync: $8/month — end-to-end encrypted cloud sync across all your devices. Optional, but worth it if you work on more than one machine.
  • Obsidian Publish: $16/month — turn your notes into a public website. Completely optional.

For most people, the cost of Obsidian is zero. The $8/month Sync add-on is the first meaningful upgrade, and only if you move between devices.

Notion — Your Shared, Structured Hub

Notion is a workspace platform that handles structured information: databases, project boards, team wikis, and linked documents. Where Obsidian excels at freeform, private thinking, Notion handles organized, linkable, shareable knowledge.

The mental model for Notion is a relational database with a word processor attached. You build pages, link them, filter them, and view them in different formats — table, calendar, kanban, gallery. Notion AI (bundled with paid plans) lets you ask questions about your pages, auto-summarize long documents, and generate drafts from bullet points.

Notion Pricing:

  • Free: Unlimited pages and blocks for individuals. Limited file uploads (5MB), no page history beyond 7 days, no advanced permission controls. Enough to get started.
  • Plus: ~$12/month (billed annually) or $16/month (monthly) — unlimited file uploads, 30-day page history, unlimited guests. This is where Notion AI is also included.
  • Business: ~$18/month — SAML SSO, advanced analytics, private team spaces. Only relevant at team scale.

For a solo PKB, the free tier is a real starting point. The Plus plan at $12/month (annual) is the logical first upgrade — mainly because Notion AI becomes useful fast once you have a dense enough knowledge base to ask questions about.

Claude — The Intelligence Layer

Claude is Anthropic's AI assistant. In the context of a PKB, it's not a storage tool — it's a thinking partner. You use it to process raw notes into structured summaries, to connect ideas across documents you paste in, to write first drafts from research notes, and to ask questions about content you're working with.

The key capability for PKB use is Claude's long context window: it can read and reason about very large documents — dozens of pages at once — without losing track of details from the beginning. This makes it genuinely useful for synthesizing material, not just answering short questions.

Claude Pricing:

  • Free: Available with rate limits. Enough to test the workflow and get real value on lighter use.
  • Claude Pro: $20/month — priority access, significantly higher usage limits, access to the most capable models. Worth it once you're feeding Claude long documents regularly.

Total monthly cost at the lowest paid tier: $0 (free versions of all three) to ~$40/month (Obsidian Sync + Notion Plus + Claude Pro). Most people start free on all three and upgrade the one they actually hit limits on first.

How the Three Layers Work Together

Think of your PKB as a three-layer stack, each with a clear job:

  • Capture layer (Obsidian) — everything you read, think, learn, and draft goes here first. Raw, fast, zero friction. No organizing required at input time.
  • Structure layer (Notion) — organized, shareable, project-linked knowledge. Wikis, databases, dashboards. The source of truth for anything that needs to be navigated or shared.
  • Intelligence layer (Claude) — the processing engine. You bring raw material to Claude, ask it to synthesize, rewrite, connect, or expand, and push the cleaned-up output back to Obsidian or Notion.

The workflow isn't a one-time setup. It's a loop: capture → process → organize → reference — repeated indefinitely. Each cycle adds more material and more connections, which makes the whole system more valuable over time.

The Full Workflow, Step by Step

Step 1: Capture in Obsidian

When you read something interesting — an article, a book chapter, a podcast takeaway — open Obsidian and create a note. Don't organize it yet. Use the Daily Notes plugin to create a note for today and just write. Add a tag. Link to another note you've already written using [[double brackets]] if you remember one that's related.

The only rule at capture time: get it down before it disappears. Speed matters more than structure here. You can always promote a rough note into a permanent file later. A note that doesn't exist is useless; a messy note that exists is recoverable.

Step 2: Process with Claude

When you've been reading about a topic for a while — a week of notes, a book summary, a cluster of linked ideas — bring that material to Claude. Copy and paste your raw notes into a conversation and ask Claude to do something specific with them:

  • "Summarize the core arguments across these notes. Flag any contradictions."
  • "I've been reading about [topic]. Write a 400-word synthesis from this source material."
  • "What gaps or questions are missing from this body of notes?"
  • "Rewrite these bullet points as a structured article outline."

This is the step that turns a PKB from a storage cabinet into a thinking tool. Claude doesn't just retrieve your notes — it reasons about them, finds patterns you missed, and produces output you can actually use.

If you use Claude's Projects feature (available on Pro), you can upload reference documents that Claude keeps in context across multiple conversations — useful for long-running research where you don't want to re-paste the same background each time.

Step 3: Organize in Notion

The synthesized output — the cleaned-up summary, the project page, the structured wiki entry — goes into Notion. This is the permanent, organized layer: well-titled pages, logical categories, linked databases.

Once your knowledge is in Notion with enough density, Notion AI becomes useful for a different kind of retrieval. Instead of searching for a page by name, you can ask "what do I know about X?" and get an answer that draws from your entire workspace. It's not perfect, but it's genuinely faster than manual search when you have hundreds of pages.

Step 4: Reference and Repeat

When you start a new project, you search Obsidian for related notes from past work, bring the relevant context into Claude, and use Claude to connect old thinking to new problems. The output goes into Notion alongside the project. Over time, your PKB compounds — each new thing you learn is in dialogue with everything you've already captured.

This compounding effect is the whole point. A PKB that's six months old is dramatically more useful than one that's a week old. The value isn't in any single note — it's in the accumulated density of connected ideas.

Practical Setup: The Minimal Version That Actually Works

Most people stall on the setup. They spend the afternoon designing the perfect folder structure and never actually start taking notes. Here's the minimal system that avoids that trap:

In Obsidian:

  • Four folders: Inbox / Notes / Areas / Archive
  • Enable the Daily Notes plugin. Every morning, open today's note. Everything goes there first.
  • Use tags, not nested folders: #reading, #ideas, #project-name. Tags are searchable and don't require you to decide where something belongs at input time.
  • Once a week, spend 15 minutes promoting notes from Inbox to Notes or Areas. That's the only maintenance required.

In Notion:

  • One database called "Knowledge Base" with properties: Title, Category, Status, Date.
  • One page per topic or project — not per note. Notion pages should be substantial, not a sentence each.
  • Use Notion AI to summarize when pages get long. Keep pages navigable.

With Claude:

  • Keep a prompt you reuse: "Here are my raw notes on [topic]. Summarize the core ideas, flag any gaps or contradictions, and write a 3-paragraph synthesis I can paste into my knowledge base."
  • Paste notes, get output, copy what's useful back into Obsidian or Notion.
  • That's the whole workflow. It works on the first day and it scales to years of accumulated material.

What Each Tool Does Badly (And How to Compensate)

No tool is perfect. Knowing the weaknesses in advance saves you from discovering them the hard way:

Obsidian has no native AI unless you install community plugins. Its mobile app is functional but not as fast as dedicated mobile note apps. And the graph view, while impressive-looking, becomes visually overwhelming once you have more than a few hundred notes — useful for exploration, not for navigation. Fix: don't rely on the graph for navigation. Use search and tags instead.

Notion has notoriously slow search at scale. It can feel sluggish on large databases, and the free tier's file upload limit is a real constraint if you're working with images or PDFs. Fix: use Notion for organized output only — never as a capture tool. Keep it lean enough that you know what's there.

Claude has no persistent memory of your notes unless you explicitly provide context in each conversation. Fix: paste what's relevant rather than expecting Claude to remember a previous session. For recurring projects, use the Projects feature with uploaded files to maintain context without re-pasting every time.

Mistakes to Avoid

These patterns are consistent across people who set up a PKB, get excited, and then abandon it two months later:

  • Building the system instead of using it. The perfect folder hierarchy is not a PKB. Notes are a PKB. Start capturing before you finish designing.
  • Using all three tools for the same job. They have distinct roles. Obsidian is not Notion. Notion is not Claude. Claude is not either. Respect the separation.
  • Waiting until you have "enough" notes to use Claude. Start early. Ten rough notes on a topic can produce a useful synthesis. You don't need a hundred before Claude becomes valuable.
  • Over-engineering Notion. Twenty properties in your database that you never fill in is visual debt that makes the system feel like work. Start with three columns and add more only when you genuinely need them.
  • Treating the PKB as a library. The point is not to store everything you've ever read. It's to store the things you actually want to think with. Be selective at the capture stage, not the organization stage.

Is This Overkill for You?

Honestly: maybe. If you capture fewer than five ideas a week and work on one project at a time, a single tool — probably Notion alone — is enough. Adding Obsidian and Claude to a low-volume knowledge practice doesn't compound fast enough to be worth the setup cost.

But if you regularly consume information to produce something from it — writing, research, product decisions, client work, building anything over time — the three-layer system has a return that a single app doesn't. Obsidian's local-first model protects you from vendor lock-in and keeps capture frictionless. Notion's structure makes organized knowledge navigable. Claude's reasoning turns accumulated raw material into usable thought.

The system pays off most for people who think across domains: you're reading about psychology and applying it to product design, or tracking research across a long project, or maintaining expertise in a field that moves fast enough that last year's notes need to be connected to this year's context.

If that's you, this setup — even at zero cost — is one of the highest-leverage things you can do with an afternoon.

Jaime Delgado

Jaime Delgado

Product Analyst & AI early adopter

Jaime has been tracking the AI landscape since the GPT-3 era. He writes about AI capabilities, model comparisons, and practical applications for builders and founders. His daily driver is Claude inside Visual Studio Code — though he also reaches for Grok, Gemini, and ChatGPT when the question is quick and the context is light. He stays genuinely open to every AI that comes along: the landscape moves fast, and so does he. Based in Spain.

View on LinkedIn