Runneth gets sharper the more it knows about you. The brain is the place that knowledge lives. This is what it is, why it matters, and the documents that move your outcomes the most.
The brain is your private, durable layer of context. Brand, audience, product, strategy, voice. Runneth reads it before every reply. The more you save into it, the more Runneth sounds like a teammate who has been at your company for a year, not a stranger who showed up this morning.
A great creative strategist is only as good as the context they're working with. Same goes for Runneth. Without context, you get generic. With context, you get on-brand, on-strategy, ready-to-test work.
Anything durable. Brand books, customer reviews, briefs, persona docs, past winners, product specs, your team's playbooks, your tone of voice. Once it's saved, it shapes everything Runneth writes for you from that point on.
Here's what the output actually looks like when Runneth has context vs. when it doesn't.
Empty brain
You ask Runneth for 5 hooks for your new electrolyte product. It writes generic hydration hooks. "Stay hydrated all day." "The cleanest electrolyte drink you'll find." On-trend, but it could be anyone.
"Drink this and feel the difference."
Fed brain
Same ask. Runneth has your brand book (anti-Big-Sports-Drink positioning), 200 customer reviews (real word: "afternoon crash"), and your last brief (target is perimenopause women, not gym bros). It writes hooks in your customers' actual language, against your real competitor, for your real persona.
"The 3pm crash is hormonal. Not a willpower problem."
Every hook, brief, concept, ad teardown, weekly report, and email Runneth produces reads the brain first. One good brand context save makes a thousand downstream outputs better. The work you do once pays off every day after.
Ranked by how much they move the quality of Runneth's output across the workspaces we've watched. You don't need all of these on day one. The first three cover most of the lift.
Your brand book, brand kit, voice and tone guide, positioning doc, or anything that explains who you are, who you're against, and how you sound. Upload the PDFs you already have. Runneth distills them into durable brand context.
Why it matters. Every hook, concept, script, brief, and ad teardown reads this first. Without it, output is generic. With it, output sounds like you wrote it.
Competitor ad library exports, screenshots of competitor campaigns, swipe files of ads you admire, hook libraries you've built. Runneth uses these as reference points, not blueprints.
Why it matters. Helps Runneth spot the category whitespace and write concepts that don't sound like everyone else.
Persona research, ICP definitions, target audience docs, the segmentation work you've already paid an agency or researcher to do. Upload it. Runneth uses it to scope which pain, which motivator, which awareness stage every concept is written for.
Why it matters. Without persona context, Runneth writes to everyone (which is no one). With it, concepts land on a real audience.
SKU list, product fact sheets, pricing, ingredients, claims, technical specs, what's allowed in copy and what isn't. Runneth treats this as ground truth and won't invent features or break claims it has on file.
Why it matters. Stops hallucinated claims. Keeps the legal team happy. Makes product-specific work accurate the first time.
The briefs and concepts that have actually performed for you. Two or three is enough. Runneth pattern-matches against these whenever you ask for something similar.
Why it matters. Past winners are signal. They teach Runneth what "good" looks like in your account specifically, not in general.
Trustpilot, Yotpo, Shopify, Amazon, your support tool, customer interview transcripts, survey responses. Runneth mines them for real language: pain points, objections, transformation moments, the exact phrases your customers use.
Why it matters. The fastest path from "good ad" to "great ad" is using your customers' words instead of yours. Reviews give Runneth that vocabulary. Tell us where your reviews live in section 6 of the checklist and Runneth will pull them directly. No exports needed.
Page summaries, CRO audits, prior A/B test results, post-purchase survey data. Anything that explains what your landing experience says and what it converts on.
Why it matters. Keeps ad copy and landing page in sync. Surfaces message-match issues before you spend budget on a mismatch.
How your Meta, TikTok, or other paid channels are actually structured. Funnel stages, attribution setup, KPI definitions, what counts as a win on each channel.
Why it matters. Runneth reads performance data through your actual strategy, not generic defaults. "Best performing ad" means what you mean by it.
Your brief template, your concept format, your weekly report shape, your hook structure. Save the example you want Runneth to match.
Why it matters. Output comes back in your team's format, not Runneth's default.
Discovery calls, kickoff notes, strategy session transcripts, post-mortems. Anything that captures decisions made in the room.
Why it matters. Keeps the brain aligned with what was actually agreed, not what's written down in a stale doc somewhere.
This is the minimum brain. Most of the lift, almost none of the time.
No. Drop the file in chat and say what it is. Runneth handles the rest.
Upload it anyway, mention what's stale, and Runneth will weigh the rest accordingly. Something durable beats nothing.
Yes. Tell Runneth what to change or remove and it updates the saved file. Nothing is locked in.
Anyone in your workspace reads from the same brain. One save, everyone benefits.
Ask. Runneth will tell you whether it would shape future outputs or whether it's better kept as one-off context.
This is the short version of what most teams upload first. For each item, attach the file and add a sentence or two about what it is so Runneth knows how to use it. Skip what you don't have. Everything routes straight into your workspace.