The honest failure mode of startup advising isn't time. It's depth. You're on five or eight cap tables, you take a quarterly call with each founder, and the calls structurally start with twenty minutes of you re-loading the company's context — what they're working on, what they raised, what the last conversation was about, who the customer is. By the time you're actually useful, you have ten minutes left. The founder is too polite to say it, but the value of the call dropped the moment you opened with "remind me where you are."
A vault of plain markdown notes with an integrated agent fixes the structural part. Not the depth that comes from genuine engagement with the company, but the friction of holding context across many companies in narrow windows. The advising hour stops being twenty minutes of re-orientation followed by ten minutes of value. The same shape underwrites how angel investors handle deal screening and portfolio tracking and how fractional executives manage multiple engagements — different roles, identical context-juggling problem.
A vault shaped around your portfolio of advisory relationships
The shape that holds up across portfolio size is roughly: one top-level page per company you advise, with sub-pages for the running thread of conversations, the company brief, the decisions log, the metrics and updates archive, and any artifacts the founder has shared with you (decks, memos, financials).
Capy supports unlimited page nesting, so a deeper engagement — board seat, hands-on operator role, weekly working sessions — can fan out into more granular sub-pages, while a lighter advisory relationship stays simple. The whole vault is plain markdown. That matters because before each call with a founder, you ask the agent to read the company's pages and write a one-page pre-call brief. Five minutes. The call starts with you remembering the company, not asking the founder to re-orient you.
A pre-call brief that respects the founder's time
The mechanic that turns advising calls from polite-but-thin into actually-useful is the pre-call brief. A working flow: an hour before the call, ask the agent in your vault to read the running thread, the most recent updates the founder has shared, the open questions from last call, and any new artifacts. Have it draft a one-page brief: what's open, what they last raised, what you committed to think about, what's the most leveraged thing you should ask about today.
You read it. You adjust it. You walk into the call having actually remembered the company. The first thirty seconds is "I've been thinking about the question you raised in the last call about pricing — let me ask you something" instead of "remind me where you are." The founder feels the difference immediately. (The same brief-from-the-vault pattern is the spine of how account managers keep client context from slipping and annual planning — different settings, same warm-up discipline.)
Per-founder running threads that survive between calls
The hardest part of advising isn't the call itself; it's the gap between calls. Three months pass between quarterly check-ins, and in that gap you've taken five other calls with five other founders, read a dozen decks, and lost most of what was specific about this one.
A working setup: each company's page has a "running thread" sub-page where you drop a one-line entry after every interaction. Email exchange, Slack message, intro call you facilitated for them, a thought you had about their business while reading something else. Two minutes per entry. Over a year, the running thread becomes the primary-source archive of the relationship — what you've actually engaged with, not what you remember.
When the next quarterly call rolls around, the agent reading the running thread reminds you of the texture: the customer name they mentioned three months ago, the hire they were debating, the strategic question that was open. The call gets specific instead of generic.
A decisions log so your advice doesn't drift
A particular failure mode of advising is giving the founder slightly different advice on the same question across successive calls, because you've forgotten what you said last time. The founder remembers. They notice. The credibility cost compounds quietly.
A working setup: an inline decisions and recommendations database on the company page via the :::database::: directive — rows for date, topic, what the founder asked, what you recommended, the rationale, and what they decided to do. The database lives directly in the page, not in a separate tab. After every call, ask the agent to read the recap and propose entries. You confirm or edit; the rationale lands while it's still fresh.
When the same topic comes back in the next call, you read the prior entry first. Either you give the same advice (consistent, signals you've thought it through) or you give different advice with explicit acknowledgment that you've changed your mind and why (signals you're actually engaging). Both beat the random walk that comes from advising from memory.
Founder-shared documents the agent can actually read
Most founders share things with their advisors — the deck, the financial model in PDF form, the strategy memo, the customer interview notes, the candidate's resume for the head-of-X hire. Most advisors skim these once and then can't find them when the question about them comes up two months later.
Drop the documents on the company page. They auto-convert to markdown via docstrange, which means the agent can read them as searchable text the same as any other note. When the founder asks "what did you think about the section on our pricing assumptions in the strategy memo," you ask the agent to pull the relevant section and tell you what they actually said. You have an opinion based on actually reading it, not a hand-wavy memory.
The conversion runs once per upload and the document stays searchable from then on. This is what makes "chat with the deck" actually work for an advisor who's seen twenty decks this quarter — the agent has read all of them as text. (The same workflow underwrites due diligence in acquisitions — different artifact, same mechanic.)
Cross-portfolio pattern recognition, with board-session recordings alongside
The advisor's structural advantage is the cross-portfolio view: you've seen this kind of company before, you've watched a similar founder navigate the same decision a year ago. Most advisors carry that pattern recognition in their head. The bandwidth limit means you only access it for the most-recent, most-vivid examples.
A working flow: when a founder asks you a question that feels familiar, ask the agent to search across your other companies' pages for any decisions, recommendations, or notes that touched on the same topic. The agent finds the prior cases and quotes them. You're not relying on memory; you're reading what you actually saw and recommended.
This is the advising equivalent of an investor's pattern-matching, except it works on the actual texture of your portfolio rather than the abstracted lessons. The same cross-portfolio search underwrites agency owners taking on more work without hiring — different setting, identical leverage.
For the heavier rooms — the strategic offsite, the executive review, the budget conversation that runs long — record the session in Capy where the company's protocol allows. The transcript comes back with speaker diarization (labels like "Speaker 1: …") so you can tell who said what. Ask the agent to draft a recap with three sections: what was decided, what was raised but not resolved, and what the dynamic in the room suggested about where the company's center of gravity actually is. The transcript matters most for the next conversation; when the company comes back two months later with a different framing of a decision the board ratified, you find the moment in the vault with the speaker labeled.
What this isn't
Capy isn't a portfolio management system or a board portal. The structured side — official board packets, the company's data room, the cap table — still lives in the tools the company uses. Capy is for the unstructured side: your personal connective layer, the running threads, the decisions log, the per-founder context. That's the part that's currently sprawled across email and Drive and your memory.
It's also single-user by design. One advisor, one vault. If your model is a multi-person advisory firm where partners need to edit the same artifacts with role-based permissions, that isn't this product. The shape that fits is the individual advisor or board member running their personal cross-portfolio context layer alongside whatever shared infrastructure the firm or boards use. Pricing tiers are on the pricing page.
A note on confidentiality: the things founders share with you are theirs. The vault is on our cloud servers and is single-user, but treat it the same way you'd treat any other workspace where you hold sensitive material — with care about what you upload and an awareness that it's not an air-gapped notebook. Most advisors are fine with the cloud-vault model; some aren't, and that's a personal call.
A small first test
Pick the company on your roster you spend the least time with. Drop the last deck they shared, the most recent founder update, and your notes from the last call into a page. Ask the agent to draft a one-page pre-call brief. If the brief catches a thread you'd otherwise have walked into the call having forgotten, that's the agent doing for you what advising at portfolio scale actually requires.
Try Docapybara free. Load one company's last quarter and see how the next call opens.