Running a nonprofit means juggling programs, donors, the board, staff, and grants on a structurally thin operational layer. There's no chief of staff. There's often no full-time fundraiser. The development director is also writing the annual report, the program manager is also writing the grant report, and you — the executive director — are the connective tissue holding the whole thing together. The week feels permanently overcommitted because it is.

A vault of plain markdown notes with an integrated agent isn't going to add staff capacity, but it is going to make the connective work tractable in the time you actually have. Not a CRM, not a project tool, not a donor database — a calm working layer that holds the context across programs and people, so the writing, the prep, and the synthesis stop sprawling. The same shape underwrites [grant writing and applications](/guides/founders-ceos/grant-writing-applications/) and [AI for fundraising work](/blog/ai-for-fundraising/) — both of which intersect heavily with what the ED is doing every week.

## A vault shaped around how the org actually moves

The shape that holds up across org size is roughly: one top-level page for the organization, with sub-pages for each program, each major donor relationship, the board calendar, the staff context, and the grant pipeline. Capy supports unlimited page nesting, so a heavier program area can fan out into per-cohort or per-site sub-pages without forcing structure on a smaller one.

The whole vault is plain markdown. That matters because when you sit down on Monday morning and try to figure out what you actually need to be on top of this week, you ask the agent to read across every page and write a one-paragraph orientation: what's open across programs, what's owed to donors, what's coming up with the board, what grant deadlines are in the next two weeks. The week goes from "all of it, all at once" to "here's the actual list."

## Program updates that don't get rewritten three times

A typical nonprofit ED writes the same program update at least three times in a year — once for the board, once for the major-donor newsletter, once for the annual report. The content is mostly the same; the framing differs. Most EDs rewrite from scratch each time because the original is in a Google Doc somewhere, the framing is hard to remember, and starting blank feels easier than reverse-engineering an old version.

A working flow: each program has a "running narrative" sub-page where the program manager (or you) drops short updates as work happens — milestones hit, stories from the field, numbers that landed. Two minutes per entry. By the time the board update is due, the narrative is half-written. Ask the agent to read the running narrative for the period and draft a board-ready one-pager. By the time the annual report is due, ask it to read the same narrative across the whole year and draft a report-ready section. Same source material, different output, no rewriting from blank.

You edit. The draft is usually 70% of the way there because the agent has read your actual program notes, not generated from generic web data. (The drafting-from-vault pattern is the same one underwriting [investor-relations and board updates](/guides/founders-ceos/investor-relations-board-updates/) — different sector, same speed-up.)

## Donor context that survives the gap between asks

The hardest part of donor work for an under-resourced ED isn't the ask itself. It's holding the context for fifty or a hundred meaningful donor relationships across a year, so each conversation feels personal instead of transactional. Without a system, the context lives in your head, and your head is currently full.

A working setup: a sub-page per significant donor with their giving history, the conversations you've had, the topics they've asked about, the family members or interests they've mentioned, and the program areas they care most about. After every interaction — call, email, lunch, gala check-in — drop a one-line entry. Two minutes. Over a year, the page becomes a primary-source archive of the relationship.

Before your next conversation with that donor, ask the agent to read the page and draft a short brief: what they last raised, what's open, what they care about, what's new in the program area they fund. You walk into the conversation having actually remembered the relationship. (The same context-holding habit is the spine of [how account managers keep client context from slipping](/guides/sales-accounts/account-managers-ai-notes-client-context/).)

## Board prep that fits in the evenings you have

The board packet doesn't write itself, and the time to write it always lands in the worst week. The fix isn't a fancier template; it's holding the source material in one place so the assembly stops being a scavenger hunt.

A working flow: by the time you sit down to draft the board packet, the agent has access to the program running narratives, the financial commentary you've kept current, the open board questions from last meeting, and the strategic-plan progress notes you drop in monthly. Ask the agent to draft a v1 of the packet by reading across all of it and following the format of last quarter's packet. Edit. Send. (The compounding board-prep mechanic shows up in our writeup of [AI notes for nonprofit board members on governance and fiduciary duties](/guides/founders-ceos/nonprofit-board-members-governance/) — board side and exec side of the same room.)

The first quarter you do this is roughly the same time investment as before. By the third quarter, the packet writes itself in an evening because the source material is already in the vault and the agent has learned the shape from the prior packets.

## Grant pipeline as a database, not a spreadsheet you forget

Most small nonprofits track grants in a spreadsheet that drifts out of date by the second month. The deadlines are right; everything else (the contact at the foundation, the prior-year application, the rationale for the ask amount, the program officer's last email) lives somewhere else.

A working alternative: an inline grants database in the org's vault via the `:::database:::` directive — rows for foundation, program officer, ask amount, deadline, status, prior-year application page, and rationale. The database lives directly inside the page, not in a separate tab. Each row links to a sub-page with the deeper context: the prior-year application as a markdown copy, the program officer's correspondence threaded by date, the alignment notes about why this foundation funds your work.

When a grant deadline approaches, ask the agent to read the linked sub-page and draft the new application using the prior-year structure with this year's program updates pulled from the running narratives. The application that used to take a week of evenings takes one long evening, and the quality goes up because the source material is genuinely current. (The grant-writing workflow is covered end-to-end in [AI notes for grant writing and applications](/guides/founders-ceos/grant-writing-applications/).)

## PDFs from foundations and consultants, finally readable

Nonprofit work generates a steady stream of PDFs you should read carefully and probably won't. Foundation guidelines (each one slightly different). Consultant reports on board governance or development capacity. The auditor's letter. The state attorney general's annual filing.

Drop the PDFs on the relevant page. They auto-convert to markdown via docstrange, which means the agent can read them as searchable text the same as any other note. Ask the agent to summarize the new foundation's funding guidelines in plain English and tell you whether your work qualifies. Or to read the auditor's letter and pull out the management-letter recommendations separately from the opinion. Or to compare two consultants' governance assessments and tell you where they agree and disagree.

The conversion runs once per upload and the document stays searchable from then on. This is what makes "chat with the foundation guidelines" actually work — the agent isn't running OCR on every query; it's reading text it already has.

The same page is where the harder conversations live as recordings — the planning retreat with the board, the difficult performance conversation with a senior staff member, the major-donor pitch where the donor's question matters as much as your answer. Record in Capy where the setting 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 we agreed, what's still open, what we agreed to revisit and when. The transcript matters less for the conversation itself and more for the next one, when the topic comes back and you need the exact wording.

## What this isn't

Capy isn't a CRM, a fundraising platform, or a program-management system. The structured side of running a nonprofit — donor records, financial system, program data, payroll — still lives where it lives. Capy is for the unstructured side, which is most of the ED's actual week: the writing, the connective tissue, the context, the draft work. That's the part that's currently sprawled across email and Drive and your memory.

It's also single-user by design. One ED, one vault. If your org needs a multi-user shared workspace with role-based permissions where staff edit the same artifacts together, that isn't this product. The shape that fits is the ED holding their own working layer alongside the org's structured tools. Capy doesn't claim regulatory certifications either — your nonprofit's compliance posture stays your responsibility. Pricing tiers are on the [pricing page](/pricing/).

## A small first test

Pick the next board meeting on your calendar. Drop the program running narratives for this quarter, the prior board recap, and the financial commentary you've been keeping. Ask the agent to draft a v1 of the board packet. If the draft catches a tension or a thread you'd otherwise have missed, that's the agent doing for you what running an under-resourced nonprofit asks for and rarely gets.

[Try Docapybara free](/accounts/signup/). Load one program area and one donor relationship and see what falls out.