Nonprofit directors are doing the work of three people and writing grants between board meetings. Claude has become a real productivity lever for this — helping draft compelling impact narratives, sharpen donor communications, and structure programme reports faster than before.
But the re-briefing problem is severe. Every Claude session requires you to explain who you are, what your organisation does, who your current funders are, what impact language they respond to, and which grants are in flight. That's ten minutes of context-setting before you can ask the actual question.
Stash gives Claude a persistent memory of your organisational context. You brief it once. It knows forever.
The knowledge that makes Claude genuinely useful for grant writing and fundraising is organisational, not generic:
Without persistent context, Claude writes generic grant prose that could apply to any organisation. With it, Claude writes from your actual impact story.
| Collection | What goes in it |
|---|---|
org | Mission, theory of change, what makes you distinctive, key statistics |
funders | One record per funder: priorities, relationship status, what they've funded, language preferences |
metrics | Current impact numbers — beneficiaries reached, cost per outcome, programme outcomes, date of last update |
programmes | Programme descriptions, logic models, current status, what you're seeking funding for |
donors | Key individual donors, relationship notes, giving history, next contact date |
applications | Grant applications in progress or submitted — funder, amount, deadline, status |
You: "context(expand=['org', 'metrics', 'funders'])"
Claude: [loads your organisational context, current impact numbers, and funder profiles]
You: "I'm writing the Impact section of our application to Esmée Fairbairn. 500 words. Their current priority is long-term systemic change, not just service delivery. Use our actual numbers."
Claude: "Based on your impact metrics and Esmée's stated priority around systemic change: [drafts 500-word Impact section citing your beneficiary numbers, cost-per-outcome figures, and framing outcomes in terms of system-level shift rather than individual service counts — including the note from your funder profile that Esmée responds to longitudinal data over snapshot figures]"
That draft used your real numbers and knew what Esmée Fairbairn cares about this cycle — because you stored that context once.
You: "search donors for Margaret"
Claude: [retrieves] "Margaret Hollis — gave £5,000 in 2024, focused on youth employment outcomes. Last contact was your annual report letter in March. Your note says she asked about the progression rate into sustained employment specifically — she wasn't satisfied with our 3-month figure and wanted 12-month data. We now have the 12-month cohort results. No contact since March."
You: "Draft a short email re-engaging her with the 12-month results."
Claude: [drafts personalised email that references the specific question she raised, presents the 12-month data, and keeps the tone warm and direct without being transactional]
You: "update applications — National Lottery Community Fund application was declined. Feedback: good project, weak evidence of community co-design. Note for next time."
Claude: [updates the record] "Updated. That's now logged under the NLCF funder profile too — your note will show up next time you're drafting for them."
That's institutional memory that used to live in one person's head. Now it's searchable.
Nonprofit fundraising runs on deadlines. Stash gives Claude enough context to help you triage and prioritise:
You: "search applications for anything due in June"
Claude: "Two applications with June deadlines in your store: (1) Paul Hamlyn Foundation — £30k ask, programme expansion, deadline June 20th — status shows draft in progress. (2) Local Community Foundation small grant — £3k, deadline June 28th — no draft started. Paul Hamlyn is the higher priority and the further along. Want me to pull their funder profile so you can continue the draft?"
| Task | Stash + Claude | Not Stash |
|---|---|---|
| Drafting grant narrative with your real impact data | Yes — loads your metrics and funder context | Dedicated grant management software (e.g. Submittable, Blackbaud) |
| Tracking application deadlines and submissions formally | Lightweight notes only | Your grants tracker spreadsheet or CRM |
| Donor database management | Key relationship notes only | Salesforce, Raiser's Edge, or your donor CRM |
| Board reporting | Draft support with your context loaded | Formal governance records |
Stash is the professional context layer — the knowledge that makes Claude write like it knows your organisation. The formal systems stay where they are.
One of the subtler benefits: your stored context can include language preferences, phrases to avoid, and sector-specific framing. If your funders respond to "lived experience" over "service user perspectives," or if you use a specific theory-of-change framework, store it once and Claude applies it across every draft.
You: "add to org: 'Language notes — we say beneficiaries not service users, co-design not consultation, sustained outcomes not outputs. Funders respond better to systems framing than individual-change framing for our work. Avoid charity clichés.'"
Claude: [saves] "Noted. I'll apply this across all drafts when your org context is loaded."
add to org: [mission in two sentences, what makes you distinctive, 3 key impact metrics]funderscontext(expand=['org', 'funders']) before any grant draftingAdd Stash to Claude
Connector URL (Claude Settings → Integrations → Add MCP):
https://app.stashlite.com/mcp
Free tier. Google sign-in. No card required. Pricing may change; cancel anytime.