Summary

• Listings are for information only. The funder's own page is always the final word.
• Each listing links to the funder's page and shows how far, and when, we checked it.
• This is a pilot. What it lists, how it works and whether it continues may all change.

No guarantees

The site is offered as it is. Listings can be incomplete, out of date or wrong: a deadline may have moved, an amount may have changed, or a program may have closed. Do not rely on a listing alone when you decide whether, or when, to apply.

Not advice

Nothing here is legal, financial, tax or application advice. Whether you are eligible is up to each funder.

Your use of this site

• You are welcome to read, share and link to any page.
• Please do not copy the listings in bulk or republish them as your own.
• Quotes from funders' pages stay the funders' words. Their programs, names and logos belong to them.

Your information

Your profile and saved listings stay in your own browser, on your own device. They are not sent to us or to any AI service.

Sending Feedback

When you send feedback or report a problem, your message contains what you write, the page or listing it is about, and the time you sent it. It goes to the FundingSpark team's phone notifications. It is used only to fix and improve the site.

Links to other sites

Links to funders and other sites open in a new tab. We do not run those sites and are not responsible for what they say or do.

Changes

These terms are a draft and will change. They will be replaced before any public launch.

Questions and corrections

Found something wrong or out of date? Use "Report a problem" on the listing, or see the Feedback page.

How we use AI

We use AI, and we think it is time for AI to help artists, and not the other way around.

Most listings were found, read and written with AI, then checked by code against the funder's own page. A person does not review every listing before it appears here.

How a listing is made

1. Finding. Software reads funders' own lists of programs, and an AI assistant (Claude, made by Anthropic) runs web searches for more.
2. Reading. Code reads each funder's page and pulls out the facts it states: dates, amounts, fees, who can apply and what to send. We read pages politely: we follow each site's rules for automated visitors and do not copy whole websites.
3. Writing. The AI assistant writes some summaries and researches harder cases.
4. Keeping the source. Every fact we show keeps the funder's own words in our records, so it can be traced back to their page.
5. Checking. Code looks for our deadline and amount on the funder's page. Each listing shows the result. When something does not match, the listing is held back until a person looks at it.

Our goals

• Show only what a funder's page says. Never guess a program, an amount, a date or a rule.
• Link every listing to the funder's own page.
• Say clearly how sure we are, and when a listing has not been checked.

What AI gets wrong, and how we fix it

AI and code can misread a page. They have read a minimum as a maximum, taken an award's first prize amount from years ago as today's, and mistaken an article about programs for a program. When we find a mistake, we fix the listing, change the rule that caused it, and add a test so that the same mistake is caught if it comes back.

No paid placement

In this pilot, no funder or service pays to be listed, and listings are not ranked by payment.