Kindora MCP Connector
Kindora’s MCP connector lets nonprofits access its prospecting tools directly within Claude.
How a philanthropy veteran built an AI fundraising tool for nonprofits with Claude Code
Industry:
Beneficial deployments
Company size:
Small
Product:
Claude Code
Claude Platform
Location:
North America
328 nonprofits on the platform
within months of beta launch
Prototype approach raised $100,000 in year one
8 grants from mostly cold outreach for co-founder’s nonprofit
The first challenge for a small nonprofit raising money is figuring out which funders to approach. The tools that exist are often expensive and blunt, returning thousands of potential matches that a small team has no capacity to evaluate. Kindora, co-founded by Justin Steele, uses Claude to build AI-powered prospecting that evaluates funder fit the way a program officer would, filtering thousands of matches down to the few worth pursuing. Kindora’s MCP connector lets nonprofits access its prospecting tools directly within Claude.
With Claude, Kindora:
- Attracted 328 nonprofits within months of beta launch
- 2x in monthly signup growth from January through March 2026
- Moved from working prototype to launched beta in weeks after adopting Claude Code, pushing 10x the code output
- Grew out of a prospecting approach that raised $100,000 for co-founder Steele’s own nonprofit, landing eight grants through mostly cold outreach
The challenge
Kindora MCP Connector
Kindora’s MCP connector lets nonprofits access its prospecting tools directly within Claude.
Kindora MCP Connector
Kindora’s MCP connector lets nonprofits access its prospecting tools directly within Claude.
Small nonprofits can’t find the right funders
Steele spent a decade at a major tech company leading one of its largest philanthropy teams, overseeing nearly $700M in total giving across the Americas. He knows the funder side of the equation well: program officers have specific priorities, and as Steele described it, they only "want to talk to you if you’re in my area of interest." Broad outreach rarely leads anywhere.
In 2024 when Steele co-founded Outdoorithm Collective, a nonprofit connecting urban families with nature experiences, he found himself on the other side of the table with $30,000 in the bank, dwindling fast, and hundreds of families waiting to go on trips. He went looking for prospecting tools and found them either prohibitively expensive or unhelpful. One platform he paid $4,000 for returned 3,000 matches with no meaningful filtering. "Structuring unstructured information is AI’s superpower," Steele said. "How are you not using these models to make these matches?"
The solution
Nonprofits
Turn limited resources into lasting impact. Generate grant proposals, track program outcomes, and free your team to focus on serving your community.
Nonprofits
Turn limited resources into lasting impact. Generate grant proposals, track program outcomes, and free your team to focus on serving your community.
Nonprofits
Turn limited resources into lasting impact. Generate grant proposals, track program outcomes, and free your team to focus on serving your community.
From personal script to product with Claude Code
Steele decided to start building his own solution. Using Claude Sonnet to write the code, he built a Python prototype that eliminated 90% of 3,000 funder matches as poor fits, narrowing the list to about 75 worth pursuing. He and Sally, his wife and the nonprofit’s co-founder, applied to every one and landed eight grants totaling $100,000 in the first year, most through cold outreach, funding the trips families were already waiting for.
The prospecting approach worked well enough that other nonprofit leaders started asking Steele to run matches for them. He realized this could be a product for the thousands of small and mid-sized nonprofits that can’t afford enterprise prospecting tools or the staff to run them.
Steele and his co-founder Karibu Nyaggah, a business school classmate of 25 years, started building Kindora in April 2025. They hired a developer to help move the prototype toward a real product. In July 2025, Steele tried Claude Code, and the pace changed.
"It just changed everything," Steele said. "We pushed 10 times more code than before, with fewer errors." Kindora launched its beta in August, and 50 nonprofits signed up in the first week. By the end of the beta period that fall, 200 had joined the platform.
Shipping features as fast as users can request them
Kindora runs three Claude models matched to different jobs inside the live product. Claude Sonnet 4.6 powers the AI assistant, writes funder outreach, generates deep-research briefs on prospective funders, and drives the agentic deep-search feature that autonomously scores new funder matches. Haiku 4.5 handles the behind-the-scenes batch processing where speed and cost matter: classifying user intent, scoring thousands of matches during discovery, and powering the free-tier public grant critique tool. Opus 4.6 generates newsletter content where narrative quality is the priority.
Steele’s background is in chemical engineering and philanthropy, not software development. Claude Code closed that gap. Since December, he has handled all of Kindora’s development himself through Claude Code, shipping the product’s full feature set based directly on what users ask for. One of those requests came from Briana Baker, a founder in one of Kindora’s accelerator cohorts, who wanted to teach young people how to do fundraising pitches. That evening, Steele sat down with Claude Code at 9 PM. By 1:30 AM, he had a working interactive voice tool that acts as a program officer, asks challenging questions, and then uses Claude to analyze the transcript against a nonprofit sales training curriculum.
"I couldn’t sleep because I was so excited," Steele said. He tested it himself, practicing a pitch to a real foundation. The voice tool pushed back on his nonprofit being early-stage, asked about impact measurement, and flagged where he had waited too long to lean into storytelling. The next morning, he showed it to the accelerator cohort and it was ready to use.
“We pushed 10 times more code than before, with fewer errors.”
Co-founder, Kindora
The outcome
328 nonprofits and a growing roadmap
Kindora now has 328 nonprofits on its platform, with monthly signups roughly doubling from January through March 2026. The company has been accepted into three accelerators: AWS/Deloitte Social Entrepreneur Accelerator, Blackbaud Social Good Startup Program, and Camelback Ventures, with Camelback providing a $50,000 SAFE Note investment.
Steele has also started using the platform to expand beyond grant prospecting. In March, Kindora piloted an individual donor engagement feature with Outdoorithm Collective. The tool assessed who in the founders’ network to target, and Claude generated the campaign outreach, raising $101,000 in a single month.
For Steele, the broader significance is about who gets to build technology. He has a chemical engineering degree, but he had never built web applications before starting to code with AI tools a few years ago. Claude Code made it possible for someone with deep domain expertise in philanthropy to build a product that serves that community directly. "If you have a little bit of computational thinking mindset, you can do this," he said.
The social impact sector, Steele argues, has the most to gain from AI but has been the slowest to engage. Kindora’s roadmap keeps expanding based on what nonprofit users ask for, and Steele sees the current moment as urgent. "There’s such a unique opportunity right now for people who understand their problem space, who understand their communities," he said. "The people who are closest to the problem have the best solutions. This window won’t be open forever."
“The people who are closest to the problem have the best solutions. This window won’t be open forever.”
Co-founder, Kindora