Skip to content

The History of Obsidian: From Second Brain to AI Knowledge OS

Obsidian, launched in March 2020, is a local-first knowledge management application that stores notes as plain Markdown files, emphasizing data ownership and networked thinking through bidirectional links. It rapidly gained traction within the "second brain" movement, fueled by its robust plugin ecosystem and free personal use model. By 2026, Obsidian is evolving into an "AI knowledge OS," leveraging its local architecture to provide persistent context for AI agents like Claude Code, enabling intelligent read-write operations over personal knowledge graphs.

Core Understanding

Obsidian is a free, local-first knowledge management application that stores notes as plain Markdown files on the user's device, ensuring data ownership and privacy. Its core features include bidirectional links for networked thinking, a visual graph view of connections, and a vast community plugin ecosystem (2,700+ by March 2026). The app also offers Canvas for visual thinking and Bases (introduced 2025) for structured data views within notes. Originating as a side project from the outliner app Dynalist, Obsidian launched in March 2020, capitalizing on the Zettelkasten movement's interest in linked notes and the demand for local-first tools during the COVID-19 pandemic. Under CEO Steph Ango (since 2023), Obsidian adheres to principles of Privacy, Longevity ("File over app"), and Extensibility. By 2025-2026, Obsidian is transitioning into an "AI Knowledge OS," where its vault serves as persistent context for AI agents (e.g., Claude Code). This is facilitated by the CLAUDE.md protocol and the Obsidian CLI, allowing AI to read, write, and reason over the user's knowledge graph, transforming the system into an active "thinking partner."

Key Nuances

Obsidian's strength lies in its deliberate trade-offs: its commitment to privacy means no native web client, and extensibility through plugins can lead to inconsistent quality. The bootstrapped business model, while fostering community-driven innovation and avoiding investor pressure, also means core features like native AI agents and collaboration are not prioritized, often relying on plugins or external tools. While excelling in data ownership, customization, and performance due to its local-first architecture, Obsidian faces challenges in collaboration, native AI integration, and mobile startup times. The Bases feature, introduced in 2025, significantly enhances structured data capabilities, offering speed and dynamic linking that transforms backlink utility. The integration with AI agents via the CLAUDE.md file and Obsidian CLI is a critical development, turning the vault into an AI-accessible memory layer, but still largely a power-user workflow.

Open Questions

A central question is whether AI agent access will become a native, integrated feature within Obsidian's core or remain primarily a plugin-driven, bolt-on capability. Furthermore, as AI-native collaboration and automation become standard in productivity tools, it remains to be seen if Obsidian's principled stance against native collaboration and web clients will limit its broader appeal beyond solo users. How Obsidian will navigate competition from AI-native platforms like Taskade Genesis, which offer integrated AI, automation, and app-building from day one without configuration overhead, is also a key consideration for its future trajectory.

  • Zettelkasten
  • Second Brain
  • Personal Knowledge Management (PKM)
  • Markdown
  • AI Agents
  • Local-first software
  • File over app
  • Bootstrapping
  • Workspace DNA

Original Capture

The complete history of Obsidian from Dynalist side project to 6M+ users, the second brain movement, Bases, and the Claude Code + AI knowledge OS era. Updated March 2026.

In March 2020, while the world locked down, two Canadian developers quietly launched a note-taking app built on a radical premise: your notes should be plain text files that you own, linked like a personal Wikipedia, and never locked inside someone else's cloud.

That app was Obsidian. Six years later, it has 1.5 million monthly active users, an estimated $25M ARR, a plugin ecosystem rivaling VS Code, and a community that turned "second brain" from a niche concept into a mainstream movement. It also sparked an entirely new workflow: pairing local knowledge vaults with AI agents like Claude Code to create what early adopters call a "knowledge operating system."

This is the complete history of Obsidian — from its origins inside Dynalist, through the Zettelkasten explosion, to the 2026 frontier where AI agents read, write, and reason over your personal knowledge graph. We will also explore where the second brain concept goes next — and why the future may not look like a note-taking app at all. 📝

🧠 What Is Obsidian?

Obsidian is a free, local-first knowledge management application that stores your notes as plain Markdown files on your device. Unlike cloud-native tools such as Notion or Google Docs, Obsidian never requires an internet connection and never locks your data in a proprietary format.

Core features:

  • Bidirectional Links — Wiki-style [[links]] that create two-way connections between notes, enabling networked thinking

  • Graph View — A visual map of your entire knowledge network, showing clusters, orphan notes, and connection density

  • Community Plugins — 2,000+ extensions covering everything from Kanban boards to spaced repetition to AI integrations

  • Local-First Architecture — Notes live as .md files in a folder on your computer; sync is optional

  • Canvas — An infinite whiteboard for visual thinking, connecting notes, images, and web content spatially

  • Bases (2025) — Structured data views (tables, boards, calendars) that query note metadata without plugins

By March 2026, Obsidian has:

  • 1.5 million monthly active users with 22% YoY growth

  • ~$25 million estimated ARR ($2M/month reported revenue in September 2025)

  • ~$300-350 million estimated valuation — with zero external funding

  • 2,700+ community plugins in the official directory

  • 200+ community themes

  • A Discord community exceeding 130,000 members

  • A team of just 18 people — completely bootstrapped, "100% user-supported"

Obsidian's five-word philosophy, from their About page: Yours. Durable. Private. Malleable. Independent. No other major productivity tool can make all five claims simultaneously.

🥚 The Origin Story: From Dynalist to Obsidian

The Dynalist Years (2015-2019)

The Obsidian story begins not with notes, but with outlines.

In 2015, Shida Li and Erica Xu — both graduates of the University of Waterloo in Canada — launched Dynalist, an infinitely nested outliner app. Think of it as a power user's bullet-point tool: every item could expand into sub-items, which could expand further, creating a fractal document structure.

Dynalist found a loyal niche among productivity enthusiasts and GTD practitioners. Users loved the keyboard-first design, the zooming feature (focus on any branch as if it were a separate document), and the clean interface. The app gained a dedicated community — small but passionate.

But Li and Xu noticed a pattern in their user requests:

  1. Users wanted to link between outlines — connecting ideas across documents, not just nesting them

  2. Users wanted backlinks — to see which other notes pointed to the current one

  3. Users wanted graph visualization — to see how their knowledge connected spatially

  4. Users wanted local files — not cloud-locked databases

These requests pointed away from outlining and toward something fundamentally different: a tool for connected thinking.

The Insight: Notes as Nodes (2019)

By 2019, Li and Xu had been studying a growing online movement around networked note-taking. The ideas of German sociologist Niklas Luhmann — who used a physical slip-box (Zettelkasten) of 90,000 index cards to produce 70 books and 400 articles — were gaining a following among digital knowledge workers.

The movement had a central insight: folders isolate ideas; links connect them. Traditional note-taking forces you to decide where a note "belongs" at creation time. Linked notes let ideas participate in multiple contexts simultaneously, the way concepts actually work in your mind.

Simultaneously, Roam Research was generating buzz in private beta with its block-level bidirectional linking. But Roam was cloud-only, expensive ($15/month), and stored data in a proprietary format. Li and Xu saw an opening.

Their bet: Build the same networked-thinking capability, but with three non-negotiable principles:

  1. Local-first — Your files live on your machine, in plain Markdown

  2. Free for personal use — No paywall for core features

  3. Extensible — An open plugin API so the community could build what the core team couldn't

COVID Launch: Perfect Timing (March 2020)

Obsidian launched in public beta on March 30, 2020 — the same week the WHO declared COVID-19 a global pandemic.

The timing was accidental but transformative. Millions of knowledge workers were suddenly working from home, rethinking their workflows, and looking for better digital tools. The lockdown created a captive audience of people with time to learn a new system — and Obsidian, despite its beta roughness, offered something no competitor matched: complete data ownership with zero cost.

Early adopters shared their setups on YouTube, Reddit, and the Obsidian Discord. The community grew from zero to 10,000 Discord members in six months. Plugin developers started building extensions before the API was even stable.

🔧 Origins⚡ Maturity🤖 AI Knowledge OS Era2020 — Public beta (Mar 30); COVID accelerates adoption; Plugin API; 10K Discord by Sep2021 — Mobile apps (iOS + Android); Publish & Sync; 500+ plugins2015 — Dynalist launches; Shida Li & Erica Xu begin outliner development2019 — Roam Research sparks bidirectional linking; networked note-taking takes off2022 — v1.0.0 (Oct 13); Canvas whiteboard; LYT movement; 1,000+ plugins2023 — Steph Ango becomes CEO; Properties ship; 1M users; 110K Discord2024 — Tab stacking; large vault perf; 1,800+ plugins2025 — Bases database; $2M/mo revenue; AI plugins (Smart Connections, Claudian, Cortex)2026 — 1.5M MAU; ~$25M ARR; 2,700+ plugins; AI Knowledge OS accelerates

The Steph Ango Story: From Superfan to CEO (2023)

One of the most unusual leadership transitions in tech happened in February 2023, when Steph Ango — known online as @kepano — became Obsidian's CEO.

The Winamp Skins Origin. Ango's design career started at age 16, making pixel-art skins for Winamp — the iconic MP3 player of the early 2000s. His skin "Impulse" for Winamp 5 was downloaded hundreds of thousands of times, making it one of the most popular skins ever created. As he described in a recent interview: "I make this random little thing, pixel art stuff, just for fun... but then hundreds of thousands of people download it and use it. It changes your perspective on what is the world actually."

That experience — making something beautiful for yourself and discovering it resonates with a massive audience — became the template for everything Ango would build, including the Minimal theme for Obsidian.

From TiddlyWiki to Obsidian. Ango spent over a decade searching for the perfect journaling tool, falling in love with the wiki-link notation ([[double brackets]]) along the way. He spent a year customizing TiddlyWiki before discovering Obsidian in April 2020 — right as the pandemic was kicking off. He imported everything and went all-in after just two days. As he described in a video interview with Nicole van der Hoeven:

The Minimal theme won the community's "Gems of the Year" Best Theme award, which led to DMs with founders Li and Xu (known in the community as Licat and Silver). They started asking for advice on small business questions. The conversations grew naturally over two years.

When Obsidian approached its v1.0 release (October 13, 2022), Ango was asked to create the new default theme — bringing his Minimal sensibility into core. He worked part-time with the team to nail down details they considered "still kind of embarrassing" before calling it 1.0.

The CEO role emerged organically. As Ango put it: "I just realized it was so much fun and I love Obsidian so much... Shida and Erica are such mature entrepreneurs. We have very complementary skills." He praised Licat as "one of the most incredible engineers I've ever worked with" and credited Erica (Silver) as "the person who's influenced why Obsidian has such an amazing community more than anybody else."

Under Ango's leadership, Obsidian articulated three core principles that guide every product decision:

  1. Privacy — Ownership over your files, local storage, end-to-end encryption for sync

  2. Longevity — Plain text persists for centuries; apps become obsolete

  3. Extensibility — Users and developers can customize everything

The longevity principle runs deepest. Ango articulated it in his widely-cited essay "File over app": "The file is way more important than the apps. Obsidian will eventually become obsolete because technology changes so much. But hopefully, the plain text will continue to live forever." He frames this as a question about civilization itself: "How will the ideas that we have right now continue to persist for hundreds of years in the future? We've had books for thousands of years, but we've only had digital files for 70 or 80 years."

Each principle has deliberate trade-offs. Privacy means no web client (they can't see your files). Extensibility means some plugins feel "janky" — Ango calls this a conscious choice: "Obsidian has chosen to be bad at that, with the trade-off that it allows a lot of freedom for people to customize the app." Longevity means Markdown with all its limitations. As Ango explained: "Strong principles force those kinds of trade-offs where you could take either side and make a compelling and legitimate app."

The "blob" problem. In a 2025 interview, Ango drew a parallel between car design and software: just as every car has converged to the same aerodynamic "gray blob" for fuel efficiency, web design has converged to the same Bootstrap-derived patterns for usability. Obsidian's plugin system is the deliberate antithesis — a bet that creative expression in software still matters, even if it costs some consistency. It's the same philosophy that made Winamp skins matter in 2004: give people a canvas, and the ones who care will make something extraordinary.

On AI, Ango was prescient in 2023: "I would like, someday, to be able to have a plugin that can do some training on your vault and learn from it and do the inference locally on your computer, so that it completely aligns with the principles of privacy and local-first." By 2025-2026, plugins like Smart Connections, Claudian, and Cortex began delivering exactly this — though the Claude Code integration relies on external AI rather than fully local inference.

Ango brings a multi-disciplinary background unusual for a tech CEO. Before Obsidian, he co-founded Lumi (packaging supply chain), Inkodye (sun-activated fabric dye), and Slash Packaging — all Y Combinator-backed. He trained in biology and industrial design. His design contributions to Obsidian extend beyond themes: he created the Flexoki color scheme (an ink-inspired palette used across developer tools) and the open JSON Canvas file format that makes Canvas files interoperable — living proof of the "file over app" philosophy.

His broader design philosophy — obsessing over every detail across mediums, from software to furniture to food — extends to a side project called Computer Show, where he pairs people with computers that match their personality. As he described in a recent interview: "What happens when you consider the maximum number of details? How does it change how you perceive that thing? How does it change the experience people have when they interact with that object or software or dish?"

Under Ango's leadership, Obsidian shipped Canvas, Properties, Bases, and the CLI tool — plus his personally authored Obsidian Skills (13.9K GitHub stars) that teach AI agents how to work with Obsidian vaults.

📈 Growth: The Second Brain Explosion (2020-2023)

Why Obsidian Won the Zettelkasten Community

When Obsidian launched, several tools competed for the networked-thinking space:

Obsidian won the Zettelkasten community for three reasons:

  1. Markdown files = zero lock-in. Every note is a .md file you can open in VS Code, iA Writer, or any text editor. If Obsidian disappeared tomorrow, your data would still be intact. This resonated deeply with users burned by Evernote's decline (see our history of Evernote) and suspicious of cloud-only services.

  2. The plugin ecosystem became a moat. Because Obsidian's core was deliberately minimal, the community built the features they needed. Dataview (query notes like a database), Templater (dynamic templates), Calendar (daily notes), Excalidraw (visual note-taking) — each plugin attracted its own user community. By 2022, the plugin directory was a feature advantage no competitor could replicate.

  3. Free for personal use — forever. While Roam charged $15/month and Notion upsold AI, Obsidian's core remained free. Revenue came from optional Sync ($4-8/month), Publish ($8/month), and commercial licenses ($50/user/year). This pricing model attracted students, researchers, and individual knowledge workers who couldn't justify subscription costs for a thinking tool.

The YouTube Knowledge Ecosystem

Obsidian's growth was inseparable from YouTube. A generation of productivity creators built their channels around Obsidian workflows:

  • Nick Milo (Linking Your Thinking) — Created the ACE framework (Atlas, Calendar, Efforts) and the Ideaverse concept. ACE and Tiago Forte's PARA method (Projects, Areas, Resources, Archives) became the two dominant organizational frameworks for Obsidian vaults. His courses and YouTube videos taught hundreds of thousands of users how to structure an Obsidian vault for long-term knowledge management. His recent Home Note video showcased the Bases feature for project prioritization, daily notes, and an "On This Day" journal that resurfaces notes from the same date across years.

  • Ali Abdaal — Released a comprehensive free Obsidian tutorial (originally a $97 course) covering the full system from vault setup to advanced workflows, reaching millions of views and bringing Obsidian to a mainstream productivity audience.

  • The Zettelkasten YouTubers — Creators like Odysseas demonstrated minimalist setups focused on writing over features, warning against the "Frankenstein's monster" trap of combining too many plugins and methods. Their message: the system's power is in its simplicity, not its complexity.

This creator ecosystem produced a flywheel: YouTube tutorials → new users → community plugins → more content → more users. By 2023, searching "Obsidian setup" on YouTube returned thousands of results — more than any competing tool.

The Zettelkasten Philosophy: Why Connections Beat Folders

To understand Obsidian's appeal, you need to understand the philosophy it embodies.

Traditional note-taking is top-down: you create folders first, then file notes into categories. This mirrors how schools teach organization — subject folders, chapter subfolders, neat hierarchies. The problem: ideas don't fit in one box.

The Zettelkasten method, popularized by Sönke Ahrens's 2017 book How to Take Smart Notes, inverts this. It's bottom-up: you write atomic notes (one idea per note), link them to related notes, and let structure emerge from the connections.

Three principles that made Obsidian the ideal Zettelkasten tool:

  1. Atomic notes — Each note should contain a single idea, written in your own words. Obsidian's file-per-note architecture naturally enforces this.

  2. Bidirectional links — When you link Note A to Note B, Note B automatically shows a backlink to Note A. This creates a web of associations that grows richer with every note you add.

  3. Emergent structure — Instead of pre-defining categories, you let clusters form naturally in the graph view. Over time, dense clusters reveal your genuine interests — not the categories you thought you should care about.

The result is what Niklas Luhmann described as a "conversation partner" — a system that surprises you by surfacing connections you didn't consciously make. This is the "second brain" promise: not just storage, but augmented thinking.

🏗️ Architecture: What Makes Obsidian Different

The Vault: Just a Folder

An Obsidian vault is simply a folder on your filesystem. Inside it:

  • .md files = your notes (plain Markdown)

  • .obsidian/ folder = settings, plugin configs, themes

  • Attachments = images, PDFs, any files you embed

This architecture has profound implications:

  • Version control: You can put your vault in Git and track every change

  • Backup: Any file sync service (iCloud, Google Drive, Dropbox) works

  • Migration: Moving to another tool means moving a folder

  • AI integration: Any tool that reads files can read your vault — including Claude Code, terminal scripts, and automation tools

The Plugin API: Obsidian's Secret Weapon

Obsidian's plugin system is arguably the single biggest reason for its dominance. The API exposes:

  • Markdown processing — Plugins can extend the rendering engine

  • UI panels — Custom sidebars, modals, and views

  • File system access — Read, create, and modify notes programmatically

  • Metadata — Access frontmatter properties across the vault

  • Commands — Register keyboard shortcuts and command palette actions

Notable community plugins (2026):

This ecosystem means Obsidian can morph into almost anything: a project management tool, a writing environment, a research database, a daily journal, or a spaced repetition system — depending on which plugins you install.

Bases: Obsidian's Database Moment (2025)

In 2025, Obsidian shipped Bases — its most significant feature since Canvas, and the update that made Kepano (Obsidian's CEO) declare it could handle "tens of thousands of files without issue."

Bases brings structured data views directly into notes, without requiring the Dataview plugin:

  • Table views — Sort, filter, and group notes by any property

  • Card views — Visual galleries with thumbnails (book covers, movie posters, project cards)

  • Multiple views per base — One base can have dozens of filtered views — books, shows, quotes, people — each showing a different slice of the same data

  • Live inline editing — Edit properties directly in the table without opening the note

  • Embeddable views — Insert a base view inside any note using ![[base#view]] syntax

The three features that make Bases transformative:

  1. Speed. Bases uses heavy optimizations for scrolling large tables. Nick Milo reports instant rendering across his vault of thousands of files — a stark contrast to the Dataview plugin, which could lag on large vaults.

  2. Dynamic Links. This is the feature that changes everything. When you embed a Bases view in a note, it auto-updates based on which note you're viewing. Click on a "Peter Jackson" note, and the sidebar instantly shows every movie, journal entry, and project linked to Peter Jackson — sorted by your rating. Click on "Arnold Schwarzenegger" and it updates to show Arnold's movies, quotes from 2009 you'd forgotten about, and journal entries mentioning him. This transforms backlinks from a flat list into a three-dimensional map of your knowledge.

  3. Sidebar-embedded context. Bases views can live permanently in the sidebar. As you navigate between notes, the sidebar auto-updates to show all linked content across your entire vault — organized by folder (Atlas, Calendar, Efforts) in the ACE framework. This replaces the traditional backlinks panel with something far richer: you see not just what links to this note, but where those links live and what they mean.

How Nick Milo uses Bases in his Ideaverse Pro workflow:

  1. Project prioritization — A Bases table ranks active, simmering, and sleeping projects by a numeric rank property. Changing a rank instantly re-sorts the view. Projects move between states via right-click → move file.

  2. Daily note calendar — A Bases calendar shows all time-stamped notes with linked context, making daily notes useful beyond the day they were written.

  3. "On This Day" journal — A Bases query embedded in the daily note template surfaces every note created on the same date across all previous years — replicating Day One's nostalgic feature natively.

  4. Source library — 1,600+ sources (books, movies, shows, articles) with card views showing cover images, ratings, directors, and related notes.

  5. Recent captures — A sidebar Bases view shows the most recently added notes, so nothing gets lost in a large vault.

Why Bases matters competitively: According to a Reddit user, the combination of Notion's slowness, lack of offline support, and data lock-in drove them to seek alternatives — and Bases gives Obsidian the structured data capability that was Notion's last remaining advantage. The key difference: Bases operates on local Markdown files you own, not a proprietary cloud database.

Technical details: Bases are stored as .base files (plain text, non-Markdown) alongside your notes. The magic of Dynamic Links works through the file.hasLink(this) filter syntax — a query that returns every note linking to the currently viewed note, automatically updating as you navigate. Bases uses two-layer filtering: "All views" filters apply globally across every view, while "This view" filters narrow a specific tab. The roadmap includes Kanban board views and potentially additional view types beyond the current table and card options.

AI + Bases: Nick Milo demonstrated using Claude Code to auto-populate Bases metadata — specifically, asking Claude Code to find and add Google Books cover images to 76 book notes that lacked image properties. The AI agent identified books without images, wrote a Python script, and populated the properties automatically. This is an early example of AI agents enhancing the structured data layer that Bases depends on.

🤖 The AI Knowledge OS Era (2025-2026)

From Second Brain to AI Operating System

The original second brain was passive: you captured ideas, organized them, and retrieved them manually. Every step required human effort.

In 2025-2026, a new paradigm emerged: using your Obsidian vault as persistent context for AI agents. Instead of explaining your situation, projects, and preferences in every AI conversation, you give the AI access to your structured knowledge — and it reads, reasons, and writes within your personal context.

This is not a hypothetical. It is a workflow being actively used by thousands of knowledge workers, accelerated by three converging trends:

  1. AI agents that can read and write files — Claude Code, Codex, and similar tools can navigate folder structures, read Markdown, and update files

  2. The CLAUDE.md protocol — A convention where a CLAUDE.md file at the vault root instructs AI agents on how to navigate, retrieve, and store information in the vault

  3. MCP connectors — The Model Context Protocol enables AI tools like Claude Desktop to access Obsidian vaults through standardized connectors (e.g., smithery-ai/mcp-obsidian)

  4. Skills and scheduled tasks — AI agents can now run recurring workflows (daily summaries, meeting processing, content drafting) against your vault

How Claude Code + Obsidian Works

The workflow, popularized by creators like Ben (Internet Vin) and others in early 2026, uses a simple architecture:

reads contextinstructsretrieveswrites backreferencespoints toObsidian VaultLocal Markdown FilesCLAUDE.mdNavigation InstructionsClaude Code / CoworkAI AgentContext FilesStrategy, ICP, Brand, ProjectsDaily LogsDecisions, Rules, UpdatesSkillsProcess Instructions

The five advantages of this setup:

  1. Persistent Context — Your AI agent has access to your full business context, strategy documents, project history, and personal preferences in every conversation. No more re-explaining your situation.

  2. Bidirectional Memory — Decisions, rules, and corrections made in AI conversations get logged back into the vault. The AI agent improves over time as context accumulates.

  3. Faster Skill Building — AI "skills" (saved process instructions) can point to reference files in the vault rather than embedding context in each skill. Update a file once, and every skill that references it is instantly current.

  4. Cross-Agent Portability — Because Obsidian vaults are just folders of Markdown files, the same context works with Claude Code, Claude Cowork, Codex, or any AI agent that can read files. Your context is not locked to one provider.

  5. Team-Wide Intelligence — Teams can share a vault via Obsidian Sync or Git, giving every team member's AI agent access to the same strategy documents, ICP profiles, brand guidelines, and SOPs.

The CLAUDE.md File: The Bridge Between Notes and AI

The CLAUDE.md file is the instruction layer that tells AI agents how to navigate your vault. A typical file includes:

  • Vault structure — Which folders contain what type of content

  • Routing rules — Where to save different types of information (decisions → daily logs, rules → preferences, insights → intelligence)

  • Retrieval patterns — How to find context for specific tasks (e.g., "For LinkedIn posts, check /context/icp.md and /skills/linkedin/voice.md")

  • Update protocols — How to log new information back into the vault

This convention transforms Obsidian from a personal wiki into an AI-accessible knowledge graph — a persistent memory layer that makes every AI interaction smarter than the last.

Obsidian CLI + Claude Code: The Native Workflow

Obsidian 1.12.7 (early 2026) shipped with a feature that turned the Claude+Obsidian pattern from a clever hack into a first-class workflow: the Obsidian CLI. The CLI exposes Obsidian's internal commands — file creation, link management, base creation, settings updates — as terminal-callable operations. Once installed, Claude Code can drive Obsidian directly rather than just manipulating raw Markdown files and hoping the UI catches up.

The setup is almost embarrassingly simple:

  1. Update Obsidian to 1.12.7+ to get the CLI bundled in.

  2. Install the Terminal community plugin inside Obsidian. This embeds a real shell as a sidebar pane, so the Claude Code terminal lives inside Obsidian — no second window, no context switching.

  3. Open a terminal in your vault folder and run claude. Claude Code automatically detects it is inside an Obsidian vault and recognizes the CLI as an available tool.

  4. Create a CLAUDE.md at the vault root that tells the agent it has full access to the Obsidian CLI and explains the vault's folder conventions (people, journal, topics, attachments, etc.).

  5. Add a commands/ folder with reusable slash commands like /journal-image-entry, /process-meeting-notes, /weekly-review. Each command is just a Markdown file containing a prompt — Claude Code surfaces them as native slash commands inside the terminal.

Once that scaffold is in place, the workflow becomes ergonomic in a way the older "drag your folder into Claude Desktop" pattern never was. You can paste a screenshot into the embedded terminal, hit /journal-image-entry, and Claude Code will read the image, extract the meaning, append it to today's daily note, save the image to attachments/, cross-link any new people into the people base, and update the relevant topic notes — all in one shot. If Claude Code crashes (and the embedded terminal does occasionally reload Obsidian on plugin updates), claude --resume brings you straight back into the same conversation.

A few non-obvious tricks from practitioners running this setup at scale:

  • Keep permissions on by default. Approval prompts are a useful guardrail even in a personal vault, especially if the agent is pointed at the wrong path or follows a bad prompt. Only consider -dangerously-skip-permissions in tightly scoped, disposable test vaults with a backup you can restore from. Cycle through approval modes with shift+tab when you need to loosen or tighten access.

  • Canvas files are not in the CLI. But Obsidian canvases are just JSON, so Claude can write them by hand. Tell it once how a canvas file is structured and it will create them on demand. Add the convention to CLAUDE.md so future sessions know.

  • Reinforce learnings by asking Claude to update its own CLAUDE.md. "Whenever I share a screenshot and say 'journal entry,' do X." Ten minutes of teaching turns into months of compounding workflow improvements.

  • /init is optional. When Claude Code runs in an Obsidian vault, it already knows the environment. The instruction file is more useful than the auto-generated boilerplate.

The deeper point is that this pattern collapses two long-running ideas. The Zettelkasten "conversation partner" promise — that linked notes would become a thinking tool — is finally true when the partner is a frontier LLM with full read-write access. And the local-first ethos that defined Obsidian's brand for six years stops being a tradeoff: you get cloud-grade AI capability without moving your knowledge to anyone else's servers. The vault stays on disk, the agent runs in your terminal, and the same files are readable by every future AI agent for as long as Markdown exists.

This is also why the Obsidian + Claude Code workflow has become a pattern rather than a one-off — it is the cleanest concrete answer to the question "what does an AI Knowledge OS actually look like?" For the bigger-picture take on workspace-native memory, see Workspace DNA: The Memory–Intelligence–Execution Loop.

The Thinking Partner Paradigm

Beyond productivity, some users are exploring a deeper use case: AI as a thinking partner that helps you notice patterns in your own cognition.

By giving Claude Code access to years of journal entries, project notes, and decision logs, users report that the AI can:

  • Identify recurring decision patterns they weren't conscious of

  • Surface connections between ideas from different time periods

  • Suggest which "simmering" projects are ready to become active, based on accumulated context

  • Draft content that authentically reflects their thinking style, because it has access to thousands of their own words

This is the Zettelkasten promise — a "conversation partner" that surfaces connections — but amplified by AI that can process your entire vault in seconds rather than relying on manual link-following.

⚔️ Obsidian vs. The Competition in 2026

The note-taking landscape has fractured into distinct philosophies:

Where Obsidian Excels

  • Data ownership: No other popular tool gives you the same level of control over your files

  • Plugin depth: The community ecosystem is unmatched — effectively crowd-sourced R&D

  • Customization ceiling: Power users can make Obsidian do almost anything

  • Privacy: Local-first means your notes never touch a server (without Sync)

  • Writing focus: Markdown + minimal UI = distraction-free environment

Performance: Where Obsidian Wins on Speed

One area where Obsidian's local-first architecture pays clear dividends is raw performance:

This speed matters. As Ango put it: "Even a small delay feels like it gets in the way of my productivity... being in the cloud, it always feels a little slow."

Where Obsidian Falls Short

  • No collaboration: Obsidian is a solo tool (1.5M users). Notion serves 100M+ users across 70% of Fortune 500 teams — collaboration is table stakes for most knowledge workers.

  • No native AI agents: AI features come through plugins, not core architecture — meaning inconsistent quality and no multi-step autonomous workflows

  • No automation: You can't trigger actions based on note changes, deadlines, or conditions

  • Mobile startup times: Users report 30+ second startup times on mobile devices with large vaults — a significant friction point for quick capture

  • Learning curve: The power-user orientation means beginners face a steep on-ramp, especially for Zettelkasten workflows

  • Plugin fragility: Community plugins can break on major updates, and quality varies widely

💰 The Business of Obsidian: Bootstrapped and Profitable

Obsidian is one of the rare software success stories that did it without venture capital.

The bootstrapped model is a deliberate philosophical choice. As Ango explained:

With zero investor pressure, Obsidian can stay small (18 people serving 1.5M+ users), ship features patiently, and let the community drive innovation through plugins rather than rushing to add everything to core. The team doesn't even know their exact user count — "we're so focused on privacy, we don't have analytics."

But the bootstrapped model also means Obsidian's AI story will likely remain plugin-driven rather than platform-native — a strategic bet that the community will build what's needed. The "multiplayer" collaboration feature is still on the roadmap (not expected until late 2026), and there is no web client because it would compromise the privacy principle. Whether these trade-offs hold as AI-native collaboration becomes table stakes remains the central question of Obsidian's next chapter.

🔮 The Future: Where Does the Second Brain Go Next?

The second brain concept is at an inflection point. Three possible futures are emerging:

1. The AI Agent Layer Becomes Standard

The Claude Code + Obsidian workflow currently requires manual setup: create a vault structure, write a CLAUDE.md file, configure access. In 2026, this is becoming easier — plugins like Smart Connections and community templates are lowering the barrier. But it's still a power-user workflow.

The question: Will AI agent access become a native Obsidian feature, or will it remain a bolt-on?

2. The Second Brain Becomes the Backend

Some users are already treating their Obsidian vault as a backend for multiple AI interfaces. The same vault provides context to Claude Code in the terminal, Claude Cowork on desktop, and even Codex for code tasks. The vault becomes infrastructure, not an app.

This is a profound shift. If the AI agent is the interface and the vault is the backend, then the note-taking app becomes a data layer — powerful but invisible.

3. AI-Native Tools Skip the Manual Brain Entirely

While Obsidian users are building AI context layers from the bottom up, AI-native platforms like Taskade Genesis are building from the top down — and the gap is widening.

With Taskade Genesis, the workspace is the AI backend. There's no separate vault to maintain, no CLAUDE.md to write, no file-routing to configure. You describe what you want in plain English, and Genesis builds a deployed, living application — complete with AI agents, automation workflows, and persistent workspace memory.

This is what we call Workspace DNA — a self-reinforcing loop:

feedstriggerscreates🧠 Memory(Projects)⚡ Intelligence(AI Agents)🔄 Execution(Automations)

Memory (Projects, conversations, history) feeds Intelligence (AI agents with 22+ built-in tools, custom tool creation, multi-agent collaboration, persistent memory across sessions). Intelligence triggers Execution (reliable automation workflows with branching, looping, filtering, and 100+ integrations across communication, CRM, payments, development, and more). And Execution creates new Memory — completing the loop.

The result: every interaction makes the system smarter. Every automation creates context. Every project feeds the agents. This is what Obsidian's Claude Code workflow achieves manually — but Taskade Genesis delivers it natively, for teams, from day one.

What Genesis can do that Obsidian cannot:

  • Build and deploy apps from a prompt — Describe a CRM, dashboard, or intake portal in plain English. Genesis creates a living application with UI, data, agents, and automations — published to a custom domain, password-protected, embeddable. Over 150,000 apps built and 3 million+ automations executed since launch.

  • Custom Tool Schemas — Define business-specific agent tools unique to your workflow. No other workspace platform offers this. Agents aren't just chatbots — they're specialized teammates with custom instruments.

  • Multi-agent teams — Multiple AI agents collaborating on complex tasks, each selecting the optimal model from 11+ frontier models (OpenAI GPT, Anthropic Claude, Google Gemini). Not one agent reading files — an agent team executing workflows.

  • Public agent embedding — Deploy customer-facing AI agents directly on your website. Obsidian agents are private CLI tools; Genesis agents can serve your users.

  • Enterprise-grade automation — reliable automation workflows (the same infrastructure Netflix and Uber use) with 15+ trigger types, 50+ actions, branching up to 7 paths, loops, filters, and automatic retry with exponential backoff. Integrations span Stripe, Shopify, Salesforce, HubSpot, Slack, Gmail, GitHub, and 90+ more.

  • 7 native project views — List, Board, Calendar, Table, Mind Map, Gantt, Org Chart — no plugins, no configuration.

  • Real-time collaboration — Teams edit simultaneously with 7-tier RBAC (Owner, Maintainer, Editor, Commenter, Collaborator, Participant, Viewer), built-in chat, and video calls. SAML SSO, AES-256 encryption. Obsidian has zero collaboration features.

  • 150,000+ community apps — Clone and customize from the Community Gallery, or start from 195+ curated templates.

The trade-off is honest: Obsidian offers maximum control and data ownership for solo users willing to invest in setup. Taskade Genesis offers immediate AI-native productivity with collaboration, automation, and app building — without configuration overhead. Both approaches have merit. The question is whether your second brain needs to store or execute.

🐑 Building Your AI-Powered Knowledge System with Taskade

The second brain started as a note-taking philosophy. Obsidian made it a tool. Taskade Genesis makes it a living system.

Whether you're an Obsidian power user looking for AI-native features or a newcomer who wants to skip the setup entirely, here's what Taskade Genesis delivers:

🤖 AI Agents — 22+ built-in tools with persistent memory, custom tool schemas, slash commands, and multi-agent collaboration. Agents trained on your workspace data that can research, generate, organize, and execute tasks autonomously — including while you sleep (background agents on Pro+). Choose from 11+ frontier models from OpenAI, Anthropic, and Google. Deploy customer-facing agents publicly on your website. EVE, the meta-agent, reads your entire Workspace DNA and coordinates specialized agent teams.

👁️ 7 Project Views — Every piece of knowledge rendered as List, Board, Calendar, Table, Mind Map, Gantt, or Org Chart. Switch views instantly. No plugins, no configuration.

⚡ Automation — reliable automation workflows with branching, looping, filtering, and 100+ integrations across 10 categories (Communication, Email/CRM, Payments, Development, Productivity, Content, Data/Analytics, Storage, Calendar, E-commerce). Including Salesforce, Shopify, Slack, Gmail, and GitHub.

🧬 Genesis App Builder — Describe what you want in plain English. Genesis builds a deployed, living application with UI, data storage, AI agents, and automation workflows. Publish to custom domains, password-protect, embed anywhere. The Community Gallery already has 150,000+ apps to clone and customize.

🔍 Multi-Layer Search — Full-text + semantic HNSW (1536-dim vectors) + file content OCR. Find anything across your entire workspace in milliseconds.

👥 Real-Time Collaboration — 7-tier RBAC (Owner, Maintainer, Editor, Commenter, Collaborator, Participant, Viewer), built-in chat, video calls, @mentions. Not a single-user vault — a team workspace.

Pricing: Free tier (3,000 credits), Starter $6/month, Pro $16/month (10 users included), Business $40/month, Enterprise custom.

The traction: 150,000+ apps built. 3,000,000+ automations executed. 195+ ready-to-clone templates. Since October 2025.

The vision: Code generators create files. Note-taking apps store text. Taskade Genesis creates living software — systems that remember, think, and execute. Your second brain shouldn't just store ideas. It should act on them.

👉 Build your living second brain →

📚 Further Reading

  • History of Evernote: Rise, Fall & The AI Revolution — How the original note-taking giant lost its way

  • What is Notion AI? Complete History — From Kyoto noodles to $11B valuation

  • History of Markdown — The formatting language that powers Obsidian

  • History of Mermaid.js: Diagrams as Code — From a lost Visio file to 85K GitHub stars

  • Notion vs. Obsidian vs. Taskade — Second brain app comparison

  • What is Vibe Coding? — The natural language app creation revolution

  • 12 Best AI Coding Tools — Copilot, Cursor, Claude Code & more

  • Your Second Brain, On Autopilot — How AI transforms passive knowledge into active intelligence

  • Personal Knowledge Management Guide — Building your PKM system

💬 Frequently Asked Questions About the History of Obsidian

  • Captured 2026-04-17

Source: inbox/2026-04-17-history-of-obsidian-second-brain-to-ai-knowledge-os-2026-taskade-blog.md