Omnifact

We just open-sourced a small LLM abstraction library for TypeScript that allows you to build against LLM providers in a unified way. In concept similar to the Vercel AI SDK, but just the actual abstraction layer, minimal dependencies, no ui, no agents, no bloat link below 👇

I need some help with our Microsoft for Startups situation: As a technical founder, I made the strategic decision to move our complete infrastructure from AWS to Azure last December when we got accepted into Microsoft for Startups. The program promised a clear path: use >50% of initial credits + complete verification checklist = extend sponsorship up to $150k. We went all in. Migrated everything, rebuilt our infrastructure, and launched our product in Spring. As we scaled, we integrated deeply with Azure services. Then things got weird. Microsoft silently introduced an “Engagement Score” requirement mid-program. Despite growing our usage and building extensively with their services, we’re apparently marked as “unlikely” to qualify for extension. The frustrating part? Support can’t explain how this new score works or what we need to do. Now they’ve stopped responding to our emails all together and our sponsorship expires in 2 weeks. We’ve made a strategic decision and invested significant engineering resources into building our product around Azure’s ecosystem, but what started as a promising partnership now feels uncertain. Has anyone dealt with this? I’d love to hear about your experiences with the program, especially around credit extensions. Looking for others who’ve encountered this new score, any success stories, or contacts who might be able to help navigate this. Really appreciate any insights or introductions that could get us out of this situation 🙏 @msft4startups

Do I know anyone who could get me into the "Building with AI" happy hour at @ycombinator today? Would love to attend!

To everyone who supported us today, thank you so much. This was a wild ride, but we are so thankful for all the interactions, both online and in person, we got to experience today. I didn't know what to expect from today, but getting over 300 upvotes for @Omnifact_AI means so much to us. The whole team poured their heart into this launch and we're so excited for what comes next. @RamNoww @erzaehlsalex @DesignWithAllie @phelmig Thank you, thank you, thank you!

Hey, we're trying to hold our spot on #4 on @ProductHunt, but we're close to being overtaken. If you haven't yet, please give us your support, it would really mean the world to us. Also please share with anyone who might be interested! Thanks so much 🙏🏻 producthunt.com/posts/omnifa…

Our SF team is going to be at the @ollama open-source AI demo night at @SHACK15sf on the tail-end of our PH launch day for @Omnifact_AI. Catch us there if you want to discuss getting open-source LLMs into production and celebrate our launch 🎉 Who else is going to be there?

Here we go! @Omnifact_AI is live on Product Hunt and we need your support. Our mission is to enable companies to run generative AI securely and privately. producthunt.com/posts/omnifa… It would mean the world to us if you could retweet this, so we can reach a broader audience. Thanks for all the love and support we have already received! #buildingpublic #producthunt #producthuntlaunch

Here we go, our self-signup free plan is online. This was a labor of love for the whole team and we are excited to finally be able to share @Omnifact_AI with everyone. Only a couple of hours now until our Product Hunt launch, we're so excited. Fingers crossed that everything runs smoothly🤞🏻

I've built a small CLI tool to experiment with and test different RAG approaches and workflows locally. Currently supports the Mistral models through their API, but could be extended to support more models. Would that be useful to anyone else? Should I open source it?

I shared my thoughts on using ChatGPT in the workplace on LinkedIn, let me know what you think linkedin.com/pulse/exchangin…