Agentic development is moving faster than most of us can keep up. I’m giving a lightning talk about “The Evolution of LLM Development Powering Figma Make” at an MCP meetup hosted by @figma and @trymightyai Come join us! (Link below)

39 startups are building alongside Vercel for six weeks. Backed by $8M in credits from AWS, Anthropic, OpenAI, ElevenLabs, and more, they have the resources to scale their infrastructure. Meet the 2026 Vercel AI Accelerator cohort. vercel.com/blog/2026-vercel-…

Honored to be selected by @vercel as part of their accelerator pushing AI safety and security forward. The timing couldn't be more critical. Adversarial attacks against AI systems are getting easier to launch and harder to catch. Hidden instructions buried in PDFs. Malicious prompts disguised across images and audio. Attacks that are invisible to human eyes but perfectly visible to the models processing them. And most protection layers today? Still text-only. Still policy-only. That's the gap we built Mighty to close. We are built for the AI-native future. Citadel is a multimodal runtime security gateway. One line of integration. Sub-100ms latency. We inspect both inputs and outputs across text, images, PDFs, and audio to stop prompt injection, data exfiltration, and AI fraud before it hits your model or leaks out the other side. We're 10 to 100x faster than current state-of-the-art benchmarks and exceed them in accuracy. Whether you're running yesterday's model or tomorrow's, protection stays consistent. As AI workflows go multi-model and multimodal in production, having a security layer that is fast, simple, accurate, and actually covers the full attack surface matters more than ever. We designed one-line integration intentionally so teams can ship to production without friction. We're stopping AI fraud at runtime. Give us a try!

We red-teamed today’s open-source LLMs under real-world behavior. What we found was uncomfortable. Even the strongest models break once they leave clean prompts and enter production.

Open source is winning. The agentic future needs an immune system. For Developers who need more, our multimodal API with simple, usage-based pricing is available too: trymighty.ai/