Obsidian Second Brain: Project Notes With Clear Sources

An Obsidian second brain is a collection of notes you organize around useful questions, projects, and sources so you can recover what you learned. Start with a small vault, one home note per active project, and links that explain why each source matters. For work involving several AI assistants, fenbs can keep the selected decisions and handoffs outside your personal note collection.

How do you set up the first working vault?

Begin with a project you expect to work on again tomorrow. A real question gives the system a purpose and makes it easier to tell which notes deserve attention. Avoid moving your entire archive before you have tested how you will retrieve one useful answer.

1. Create a vault for your work notes. Obsidian defines a vault as a folder on your local file system; its vault setup guide↗ (opens in a new tab) explains creating one or opening an existing folder.
2. Add four folders named Inbox, Projects, References, and Archive. These are suggested names for this workflow, not special Obsidian features.
3. Create a project home note with the intended outcome, current question, next action, and links to supporting notes.
4. Add two real sources and write a short explanation of what each contributes. Separate your interpretation from anything the source actually says.
5. Close the project and return later. Try to find the source behind your last conclusion without relying on memory.

You do not need a plugin collection to pass that first test. If a folder name causes hesitation, rename it before adding more material. The goal is a dependable return path to your work, not an elaborate filing scheme.

What belongs in a project home note?

Keep the home note short enough to read before a work session. It should explain what you are trying to produce, which constraints matter, and what you last established. Put detailed reading notes elsewhere and link to them from the relevant sentence.

• Outcome: the concrete thing you are preparing, such as an onboarding guide for a client.
• Current position: what is drafted, reviewed, uncertain, or waiting on someone.
• Evidence: the notes and original sources behind the statements you expect to reuse.
• Next action: one step that another person could understand without reading your whole archive.

Say you are researching a new customer welcome sequence. Keep the interview notes separate from the proposed sequence. The home note might say that customers need a clearer first action, link to the interview evidence, and name the draft still awaiting review. That is more useful than a folder full of unnamed transcripts.

When the next session happens in another AI app, the same short summary can become selected project context in fenbs, with the source notes staying in Obsidian.

How should you link notes without creating clutter?

Add a link when it helps answer a question, support a statement, or find the next useful note. In Obsidian, type two opening square brackets and select a note; the internal links guide↗ (opens in a new tab) also explains Markdown links and links to headings. Use descriptive note titles so the destination is understandable before you open it.

Write a sentence around each important link. “This interview explains why new customers miss the setup email” tells you what to expect. A long list of related notes leaves you to repeat that reasoning every time you return.

Keep a source note distinct from a conclusion note. The source note records where material came from; the conclusion note states your interpretation and links back to the evidence. You can change your interpretation later without silently rewriting the original record.

What should an AI assistant receive from the vault?

Give the assistant a focused packet: the question, your project summary, relevant excerpts you are allowed to share, and clear source labels. A link or file path alone does not mean an assistant can open the content. Check what the specific app can access and provide the necessary material through its supported method.

Ask for an answer that distinguishes evidence, inference, and missing information. For the welcome sequence, the assistant could suggest wording and explain which interview notes support it. You should still open the original sources before accepting a factual claim or approving the final text.

The broader AI second brain guide explains this distinction between stored information and usable context. fenbs does not sync your Obsidian vault or automatically read its files; move selected work summaries into it deliberately.

How do you keep the system useful after the first week?

Review the notes for active work before reorganizing your entire vault. Empty the Inbox by turning useful material into named notes, attaching it to a project, or removing it when it has no purpose. A note can be interesting without belonging in the current workflow.

Read each project home note as if you had been away. Replace vague statements such as “research mostly done” with the evidence collected and the unresolved question. Mark an old conclusion as replaced when new evidence changes it, and link to the current explanation.

The review is also the moment to update an AI handoff. Record what was actually checked and what remains a suggestion. Keeping that distinction in fenbs project context helps the next connected assistant avoid treating a polished draft as an approved decision.

How to do this with fenbs

1. Create a free account and add the real project your Obsidian home note describes.
2. Manually save a concise project context note with the outcome, relevant source references, and any uncertainty. Keep private notes and unnecessary personal information out of the summary.
3. Record choices a person has approved in Decisions and rules. Add the remaining work as tasks and include what needs checking.
4. Connect a supported assistant using the integration guides. Ask it to read the project context before working and save Where we left off before ending the session.

fenbs is free for one person, no credit card. It is not the right fit if your only goal is a private, local collection of reading notes and you do not need a separate record of ongoing AI work. Obsidian can remain the main place where you think and write.

How can you tell whether the setup is working?

Try a small retrieval test: find the source for one claim, explain why you accepted it, and name the next action. If that takes a long search, improve the relevant project note before adding another layer of organization. A useful second brain shortens the path from a question to a defensible answer.

For the overall concept, read what a second brain is. The AI brain overview covers the outside work record, while Notion MCP vs Obsidian covers the narrower question of assistant access. Explore the second-brain library when you need another workflow.