The archive · Developer & Business Tools · Product decision · 2025–2026
Dedalus Labs bets 'Vercel for AI agents'; $11M seed, YC S25
YC S25's Dedalus Labs turns agent building into 5 lines of code and 3-click MCP deploys; $11M seed from Kindred and Saga.
Dedalus Labs
What the business is
An infrastructure platform that lets developers build production-ready AI agents in five lines of code and deploy MCP servers in three clicks, across any model provider.
Starting capital：$11M seed (Oct 2025) led by Kindred Ventures and Saga Ventures.
How it started
Catherine Di (CEO) and Windsor Nguyen (CTO) founded Dedalus Labs and joined YC S25 in 2025, betting that every developer hitting production-agent pain — Docker configuration, vendor lock-in, brittle integrations — would adopt a simpler infrastructure layer.
What happened
At S25 Demo Day (Sept 2025) TechCrunch's VC roundup flagged Dedalus as 'Vercel for AI agents' and one of the nine most sought-after startups in the batch. In October 2025 it announced an $11M seed led by Kindred Ventures and Saga Ventures, with angels including Hugging Face co-founder Thomas Wolf, Slack CTO Cal Henderson and Supabase CTO Ant Wilson.
How it ended up
Still building: $11M seed in hand (Oct 2025), expanding platform capabilities including an MCP Authorization Server and scaling developer adoption.
Background
Dedalus Labs is an AI-agent infrastructure platform that lets developers build production-ready agents in five lines of code and deploy MCP (Model Context Protocol) servers in three clicks. Its bet is that as models become commoditized, the value in AI shifts to the infrastructure and orchestration layer — the 'Vercel for AI agents'.
Founders Catherine Di and Windsor Nguyen joined YC S25 in 2025, targeting the developer pain of building agents: hundreds of lines of code, days of Docker configuration, vendor lock-in and brittle integrations. The platform's core insight is that agents are just orchestrations of models and tools, so making orchestration simple — with native MCP support, vendor-agnostic streaming and 3-click deployment — becomes the wedge.
At S25 Demo Day in September 2025, TechCrunch named Dedalus among the nine most sought-after startups in the batch. A month later it announced an $11M seed led by Kindred Ventures and Saga Ventures, with angels including Hugging Face co-founder Thomas Wolf, Slack CTO Cal Henderson and Supabase CTO Ant Wilson. The company is using the capital to expand platform features, including an MCP Authorization Server, and to grow its developer community.
What has to be true
- Building production AI agents today requires hundreds of lines of code and days of Docker setup — a five-line SDK is a 10x improvement developers feel immediately.
- MCP's emergence as the standard for AI tool communication creates a protocol moment for new infrastructure players to capture.
- Vendor-agnostic architecture means model upgrades don't force rewrites, lowering switching costs against incumbents.
- The investor lineup (Hugging Face, Slack, Supabase founders) signals that agent infrastructure, not another wrapper app, is where the value is moving.
What can be applied
When a protocol (MCP) is becoming the industry standard, infrastructure startups can win by making adoption trivial — 3-click deploys instead of days of setup — before incumbents standardize it.
Aftermath
As of September 2026 Dedalus Labs is post-$11M-seed and still early-stage, building out MCP-native infrastructure, an authorization server and community initiatives while scaling developer adoption of its agent platform.
Sources
- The 9 most sought-after startups from YC Demo Day
- Infrastructure Investment: Dedalus Labs Raises $11M to Democratize AI Agent Development
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card