VCRM - An Open-Source Backend for Any VC
Previously, I wrote about building an internal CRM for SAIF using Claude Code and how it replaced our patchwork of Google Sheets and Docs. A number of people reached out asking if they could see the code or adapt it for their own fund, so we decided to open-source it.
VCRM is an open-source backend template built for venture funds. You can fork it, run a single setup command, and get a working CRM connected to your own database in minutes. It handles deal flow intake, partner voting, portfolio tracking, meeting notes, internal task management, and more. The whole system is configurable from a single file, so funds can turn modules on or off depending on how they actually work. If your fund does not run formal voting rounds, you can disable that module. The navigation updates automatically based on what you enable.
Before today’s coding agents, if someone told me they needed a CRM, I would have pointed them to a SaaS product that gets you most of the way there and charges you every month forever. Custom internal software felt out of reach unless you had money to burn, engineers, or both. A small fund with a unique workflow used to have two options: live inside spreadsheets and workarounds, or pay enterprise prices for software that only sort of fit. Now there is a third option. You can build something around your actual process. This project was built specifically for SAIF’s workflow, but anyone can take the code and shape it to their own needs.
I keep coming back to the same tension I raised in my last post. I am shipping production software that I did not write and do not fully understand line by line. That is a new kind of trust relationship with software, and I don’t think we have fully reckoned with what it means yet. That is part of why open-sourcing this felt important. It is practical, because other funds can use it and adapt it but it is also about trust. If people are going to rely on software like this to track real investments and relationships, they should be able to inspect what they are using and understand how it was built. This was built with Claude Code. We reviewed it, tested it, cleaned it up, and we use it every day at SAIF. But I did not hand-write it, and pretending otherwise would be dishonest.
A few months ago, I would have assumed software like this required a startup budget or a real engineering team. Now I think a lot more people should question that assumption, because it is no longer just about this one project. VCRM is one small example of a much bigger change in who can build software, how quickly they can do it, and how closely that software can match the real needs of the people using it.
Let me know if you use VCRM and find it useful (or not). I’m also happy to help with any problems that you find installing / using the software. There are certainly still bugs and, well, that’s why it’s open sourced - so everyone can improve the product. You can find me at nick@saif.vc.
Really resonated with this. I had to build a custom admin system for my startup because our workflow is genuinely unique — how we do GTM outreach, how we aggregate community signals, how we route data between systems. No existing CRM or ops tool could handle it without bending the process to fit the software instead of the other way around. I think that's the real unlock here: small teams can now build internal tooling that matches exactly how they work, then double down on what's actually converting. The edge for early-stage startups isn't just the product anymore, it's the ability to rapidly build custom pipelines around their specific workflow. Feels like the future is something between Cursor and Retool, AI-native internal tools that shape themselves to how you actually operate, not how a SaaS vendor assumed you would, maybe thats the next billion dollar startup idea.