Bothy

Picture a development director at a nonprofit running on $400,000 a year. She is the whole development department, and grant-seeking is the hat she wears on Thursdays. Her funder list is a spreadsheet somebody built two years ago and nobody has checked since. Bothy replaces that spreadsheet with a workspace built from research a person did for her organization before she ever opened it.

A report deadline slides past on a Friday spent doing payroll

Nobody sends a rejection letter for that. A renewal she was counting on quietly does not renew, and the money is not there next year. So Bothy opens on the next thing due instead of a search box: the report due Friday, the funder she owes a reply, the application closing in nine days. The calendar sorts every grant by how soon it needs her, from Act Now (60 days or less) down to Monitor, and exports to her own calendar as an .ics file.

Somebody reads the 990s before the workspace exists

Each workspace starts from a funder list built for one organization out of IRS Form 990 filings, the IRS Business Master File, and what funders publish themselves. Twenty-five to fifty funders that plausibly give to an organization like hers, in a place like hers. Recall comes from an index of 125,548 grant lines pulled from private foundations' own filings, so the search runs on who a foundation has actually funded.

About one private foundation in eight carries no IRS subject code, so a search by subject never reaches it. Searching the grants it made does.

Finding candidates is the cheap part. Judging them is the work: when the top 40 candidates from one sweep were screened by hand, 9 survived. A funder that gets cut stays in the research tab with the reason it was removed.

Every grant gets a score, and the score shows its weights

Every opportunity is scored 0 to 5 on six criteria, and the weighted total sets its tier. A score of 4.0 or better means pursue now. The weights are sliders, and each criterion is written for that organization ("reporting load relative to current staff capacity"), not borrowed from a template.

Mission alignment

25%

Award likelihood

20%

Timeline fit

20%

Funding amount

15%

Reporting burden

15%

Match feasibility

5%

The pipeline weights dollars by stage: 5% while researching, 15% at a drafted letter of inquiry, 30% submitted, 50% pending. Eleven tabs sit around it, including a funder contact log, a reporting tracker for grants already won, and a library of reusable proposal paragraphs. A generator builds each dashboard from the research in about an hour.

65 dashboards built and given away

Every dashboard went to an organization the studio knows

None was charged. The free period was a choice: find out whether the thing is useful before asking anyone to pay for it. At least one organization has applied for a grant it found in its Bothy list and won it. The same research has since been pointed at rural hospitals and worked unmodified.

The market research in Foundation drew a clear line between what small nonprofits pay for software and what they pay for a person's work:

Grant tracker

$20–$52 a month

Funder list delivered by a researcher

$750–$3,500 once

The money follows the judgment, and the dashboard is how it gets handed over. The direction set in October follows from that: Bothy runs as a facilitated Coach House session after Thorn, where a conversation with the nonprofit's team about its goals sets the scoring weights and the backend returns the report.

It won't write the grant

That is still on her. It won't fix underfunding, and an organization with fewer than about ten active funders is fine with a spreadsheet. It isn't an AI inventing grant opportunities either. A tool that confidently surfaces a dead grant cycle costs more than no tool, which is why a person does the research and why the thing doesn't scale the way software companies need. A two-person studio can live with that.