My cofounder and I were building a company to compete with Mercor, Surge AI, etc. Scaling human intelligence, human contractors, the whole thing. A founder friend who just interviewed with YC was told by a partner: "you can't scale humans fast enough to capture the market." That stuck with me. I called my cofounder and asked one question. "What is the one place where we can access the world's info live? The internet. Not LLMs. Ask any LLM for live data and request timestamps. Their version of "live" is 24 hours to weeks old. That's not good enough.
We didn't want to build another scraper. We wanted an actual intelligence layer. Ingest everything live, before it hits the web, replicate how the brain processes signal versus noise, bring the most valuable information to the surface in real time. My cofounder is a Stanford PhD researcher in neuroscience so the potential value made sense.
The world is drowning in info, researchers/builders are overwhelmed with noise trying to find actual signal.
We did a building sprint and launched. 400+ testers in the first week. Zero customers.
I messaged every single one personally. The ones that responded, the feedback was the same: "this could be huge but setup is a pain; there's no immediate value unless you go through the steps." Got it. Users had to create an account, get an API key, manually configure MCP servers in Claude or Cursor.
The people who actually got through to it said, "Claude rated itself a 6.5 without and a 7.5 immediately with our API integration, potentially up to 9.5 with your additional agents." The signal was there. Latency sub 200ms, 5ms on follow up query's. Onboarding was killing us.
During all of this I took a moonshot and reached out to a former X AI engineer(Sully Ghori) to advise us. Rational expectation: no chance. Reality: he came on board the other night at 3am after going through everything. He came back with one sentence. "Make it all available with one click."
Last night at 1am my CTO and I made the call. Make it free. No signup. No API key. Try to lessen MCP configuration friction is still slightly there, to fix: we either have to partner with or integrate a frontier model, which needs significant funding. Agents can now onboard themselves autonomously. Humans get instant access. Enterprise stays a direct conversation to access our deep multi agent intelligence layer.
One user said, "You guys launched with a beta, if you fixed all of this its a product."
Today we're launching with a product. Try it at modernity.live. No signup, just call the API. Tell us what breaks.
My cofounder and I were building a company to compete with Mercor, Surge AI, etc. Scaling human intelligence, human contractors, the whole thing. A founder friend who just interviewed with YC was told by a partner: "you can't scale humans fast enough to capture the market." That stuck with me. I called my cofounder and asked one question. "What is the one place where we can access the world's info live? The internet. Not LLMs. Ask any LLM for live data and request timestamps. Their version of "live" is 24 hours to weeks old. That's not good enough.
We didn't want to build another scraper. We wanted an actual intelligence layer. Ingest everything live, before it hits the web, replicate how the brain processes signal versus noise, bring the most valuable information to the surface in real time. My cofounder is a Stanford PhD researcher in neuroscience so the potential value made sense.
The world is drowning in info, researchers/builders are overwhelmed with noise trying to find actual signal.
We did a building sprint and launched. 400+ testers in the first week. Zero customers.
I messaged every single one personally. The ones that responded, the feedback was the same: "this could be huge but setup is a pain; there's no immediate value unless you go through the steps." Got it. Users had to create an account, get an API key, manually configure MCP servers in Claude or Cursor.
The people who actually got through to it said, "Claude rated itself a 6.5 without and a 7.5 immediately with our API integration, potentially up to 9.5 with your additional agents." The signal was there. Latency sub 200ms, 5ms on follow up query's. Onboarding was killing us.
During all of this I took a moonshot and reached out to a former X AI engineer(Sully Ghori) to advise us. Rational expectation: no chance. Reality: he came on board the other night at 3am after going through everything. He came back with one sentence. "Make it all available with one click."
Last night at 1am my CTO and I made the call. Make it free. No signup. No API key. Try to lessen MCP configuration friction is still slightly there, to fix: we either have to partner with or integrate a frontier model, which needs significant funding. Agents can now onboard themselves autonomously. Humans get instant access. Enterprise stays a direct conversation to access our deep multi agent intelligence layer.
One user said, "You guys launched with a beta, if you fixed all of this its a product."
Today we're launching with a product. Try it at modernity.live. No signup, just call the API. Tell us what breaks.