Second Brain — a knowledge base as your operating foundation
Your judgment becomes an asset. Your AI becomes staff.
A scattered pile of files is fine — we start by deciding, together, what is worth keeping.
Consult on LINEThe information exists — you just can’t use it when you need it.
When knowledge isn’t in a form you can search and reference, everyday work stalls like this.
Searching has become the job
“Where is that file?” “How did we decide last time?” Every search stops the work. The information is in-house, but you can’t pull it out at the moment you need it.
Judgment lives only in one head
How you quote, where you draw the line on saying no — it’s all in the owner’s head. On the days that person can’t work, the company stops too.
Ask AI, and it doesn’t know your company
General AI only returns generalities. It doesn’t know your clients, history, or policies — so its drafts and calls aren’t usable as-is.
Compile the knowledge before you add tools.
Automation and quality control can only stack on a foundation. Compile scattered materials, conversations, and past decisions into a searchable “judgment asset.” Only with that asset in place does workflow redesign — and agent implementation — hold steady. Reverse the order and you just accumulate tools while the work stays the same. This is why A&A starts from the knowledge base above all else.
What is a Second Brain?
A Second Brain is a knowledge base that unifies a company’s materials, conversations, and past decisions into a form AI agents can search and reference. The term spread as a personal note-taking method; at A&A we design it as an operating foundation that stores the judgment criteria of the business itself.
Relationship to knowledge bases and RAG
Technically, it becomes the foundation for RAG (retrieval-augmented generation) — feeding internal documents to AI. But RAG only retrieves; deciding what to keep and how to structure it is human editing. A Second Brain refers to the knowledge base once that editing has made it “usable by AI.”
Collect, compile, put to work.
- 01
Collect
Contracts, meeting notes, chats, past quotes. Gather the record of daily decisions into one place, scattered as it is.
- 02
Compile
Structure it into searchable units and verify against stale or contradictory information. This is the step that needs human editing — and it matters most.
- 03
Put to work
AI agents reference the compiled knowledge base during work. Drafts, classification, and checks that fit your context run inside real operations.
A&A runs on this system itself.
This isn’t a diagram for a pitch — it’s the configuration A&A actually runs every day. We accumulate a record of decisions in a knowledge base, and the agents that reference it draft articles, diagrams, and research. This site’s articles are produced the same way.
Last updated: July 11, 2026
OSS · Public
claude-second-brain-jp
The core of the system is published as OSS — a Claude Code plugin that grows a record of decisions into a knowledge base on Obsidian. Completely free, MIT licensed, currently v0.1.0.
View on GitHub →
From free to fully handled — three ways to start.
Start free with the OSS
Want to try it yourself first? Begin with the public OSS plugin. No cost.
View on GitHub →Design alongside an advisor
Want to decide what to accumulate together? Start with the advisory plan: the first month is free, then it is ¥50,000 per month.
About advisory →Hand off the build
Want it taken from design to a running state? The upfront development fee is ¥0, followed by a fixed monthly fee that includes operations and ongoing improvement.
About the build →The upfront development fee is ¥0. After confirming the target scope on official LINE, we present the fixed monthly scope, including operations and ongoing improvement.
See how pricing works →Illustrative use cases
Where it’s likely to fit
Some are anonymized reconstructions based on real engagements; others are illustrative scenarios. Named cases will be published as permission is obtained.
FAQ
What should I accumulate first?+
Start with the judgments you use repeatedly (how you quote, where you draw the line on saying no, per-client policy) and the materials that take time to find. You don’t need to load everything at once — structure the high-frequency items first and expand while it runs.
Will internal information leak externally?+
The range of information passed to external AI services is agreed and fixed before building. Drawing the line between what’s public and what’s confidential at the design stage is part of the work. Access is also scoped to the minimum needed.
We already have files in Notion and Google Drive. Is this a rebuild?+
No rebuild. We migrate and organize what you have into a form AI can reference. Rather than rewriting from scratch, the scope is unifying scattered existing knowledge.
How is the OSS different from advisory?+
The OSS is a free tool to try yourself; advisory is working alongside you to decide what to keep and how to structure it. A tool alone doesn’t fill the editing judgment. If you want to start from “where do we begin,” advisory (first month free) fits.
Can we just add RAG and do the same thing?+
RAG retrieves materials — it’s part of the foundation. Unless a person decides what to keep and how to structure it, retrieval accuracy won’t rise. A Second Brain refers to the knowledge base once that editing has made it usable by AI.
How long until it’s usable?+
It depends on scope, but a rough guide is 2–6 weeks. We start small on a narrow target and expand while using it in real work — launching from a usable unit rather than aiming for a perfect DB up front.
A scattered pile of files is fine — we start by deciding, together, what is worth keeping, on our official LINE.
Consult on LINE- 01Add A&A on LINE
- 02A short intake chat
- 03A&A organizes the requirements
- 04A meeting only if needed