> ## Documentation Index
> Fetch the complete documentation index at: https://productbrain.io/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Your First Week with Product Brain

> A solo-user path from drifting context to a useful Chain.

<Warning>
  **Early access** — Pacing is guidance, not analytics. [**Join the Product Brain Community**](https://community.productbrain.io).
</Warning>

**Before you start:** CLI users install [Install pb](/docs/install) first (`pb --version`). MCP-only users start at [Connect your tools](/docs/integrations/overview). Read [pb vs MCP package](/docs/pb-and-mcp) if naming confuses you.

You do not need a team before Product Brain is useful.

The first useful Chain can start with one builder, one product, and one AI agent that stops trusting stale context as if it were current.

## Day 1: connect and capture one decision

Connect Product Brain to the AI tool you already use, then capture one decision you already rely on.

Do not try to document everything. Capture the one thing your next AI session must respect.

Example: "User-visible work ships behind feature flags until rollout confidence is high."

## Day 2: start a session before a work block

Before your next coding or product block, tell your agent to orient from the Chain first. Start a session so the work has a trace.

Now your agent has a place to read context from and a place to capture what changes.

## Day 3: let the agent orient before touching code

Ask the agent to use Product Brain before changing files or proposing a plan.

The change is simple: instead of relying on whichever context you remember to paste, you make orientation part of the workflow.

Notice what changes. The agent should already know the decision you captured on Day 1.

## Day 5: capture what would usually disappear

By the end of a few work blocks, capture the durable parts:

* A decision you made.
* A tension you did not resolve.
* A standard you want future work to follow.
* An insight from a user, bug, or experiment.

Ten small entries are enough for the Chain to start feeling different from a notes file.

## End of week: ask a product question

Ask your agent something it could not answer from the repo alone:

Ask: "Based on the Chain, what should I be careful about before changing onboarding?"

The point is not that Product Brain writes the perfect answer. The point is that the answer now starts from governed product context: decisions, tensions, standards, and context you already captured.

## What changes after week one

Before Product Brain, every AI session depends on whether today's context file, chat history, and prompt still match the product truth.

After one week, your Chain holds more about what changed, why it changed, and what still holds than a context file that depends on manual upkeep.

Next: learn [The Loop](/docs/method/overview), then use the [Quickstart](/docs/quickstart) if you have not connected yet.