Stop Giving Claude a Static Doc. Give It a Brain.

Every serious AI user hits the same wall within a week.

You start with a system prompt. Maybe a README. One doc dropped into a Claude Project as context. It works — until your knowledge lives across more than one file, evolves over time, or needs to travel across more than one agent.

Then the static doc breaks down fast. You are either manually updating it to stay current, or every agent you run is working with stale context.

That is the context problem. Brain solves it.

What Brain actually is

A folder in Obsidian. That is it.

Not a new tool to learn. Not a new place to write. A folder you already have, or one you create in five seconds. You work in Obsidian the way you already do. Brain watches the folder, walks every subfolder and file, and compiles everything into a single structured context note with a stable ID.

That compiled note is not just a dump of your files. It includes a manifest — every source note with its title, file path, BlackOps note ID, and a summary. The manifest is the map. Individual notes are the territory. An agent navigates with the map and drills into the territory when it needs depth. You get full coverage without blowing up your context window.

It stays current from both directions

Edit a note in Obsidian and Brain updates. Your knowledge base is always current without you touching anything in BlackOps.

But you can also go the other way. Tell an agent to add something — a new decision, a new client note, a new research finding — and the agent writes it back into the Brain via MCP. You are in Claude, or Cowork, or any MCP-compatible client, and you say "add this to the project Brain." It happens. The folder in Obsidian gets the new note. The compiled Brain reflects it.

The knowledge base is alive from both ends. You are never the bridge.

You can have as many Brains as you want

This is the part that changes how you work.

One Brain per client. One per product. One per project. One for your personal brand. One for your team's architecture decisions. One for your research library.

A consultant with eight active clients has eight Brains. Each agent loads the right one for the right job. No cross-contamination. No context soup. The right knowledge, in the right room, every time.

Every agent surface reads the same knowledge

The Brain ID goes anywhere an agent can call the BlackOps MCP. The same knowledge base travels across every surface you work from.

Claude Projects The original use case. Drop the Brain ID into your Project instructions. Every session starts with your full knowledge base already loaded. No re-explaining. No drift.

Claude Code Add get_brain to your Claude Code session startup. Your architecture decisions, brand voice, product context, and technical constraints are in the room before you write a line. Claude Code doesn't guess at your patterns. It knows them.

Scheduled tasks and automations Any agent running on a schedule — a daily summary, a content brief generator, a monitoring task — calls get_brain before it executes. Context is not a session concern. It is infrastructure. The agent knows what it needs to know before it starts.

Cowork Non-developers using Cowork to automate file and task management get the same knowledge layer you have in Claude Code. The Brain travels. The surface doesn't matter.

The mechanism is the same everywhere. One MCP call. Full context. Right knowledge for the right job.

How operators are using it now

The consultant One Obsidian folder per client. Discovery notes, architecture decisions, stakeholder context, open questions. One Brain per folder. Each Claude Project loads the right one. No copy-pasting between engagements. No stale docs.

The SaaS founder Product decisions, competitor research, customer feedback, roadmap notes — all in one Obsidian folder. Brain compiles it. Every agent conversation starts with the full picture. No re-explaining the product from scratch.

The engineering team lead Architecture decisions, RFC notes, system design docs in a shared folder. Brain compiles them into a single context the whole team references in their Claude Projects. Everyone works from the same knowledge base. Decisions don't get lost.

The researcher Papers, articles, and raw notes in Obsidian. Brain compiles a living knowledge base. Claude answers complex questions against the whole folder, pulls individual papers for depth, and writes outputs that get filed back into the vault — making the Brain smarter over time.

No RAG. No scripts. No pipeline to maintain.

There is no embedding setup. No vector database. No custom search engine to build and maintain.

You maintain your notes the way you already do. BlackOps compiles them. Agents read them.

The folder structure itself carries meaning. A note in /research/competitors/ means something different from one in /strategy/decisions/. Brain preserves that structure in the compiled output so agents understand not just what you know, but how you have organized it.

---

Brain is live now.

Sign up for BlackOps Center and create your first Brain in minutes. Point it at an Obsidian folder. Drop the compiled note ID into your Claude Project. Your agents have context from session one.