Orthogonal Raises $4.3M to Build the API Layer for AI Agents
Orthogonal (orthogonal.com), a YC W26 startup building a unified API and payments layer for AI agents, has raised $4.3M in seed funding from Pantera Capital, Pioneer Fund, Y Combinator, Decasonic, Outbound Capital, Rice Capital, Surreal by Premise, and others, as first reported by The Street.
The entire API economy was built for humans. Agents end up juggling 30+ tools, bloating context, and any API requiring identity or payment means managing accounts, API keys, billing, and auth across dozens of services. When an agent needs a capability it wasn't originally set up with, it either fails or hands the task back to a person.
Orthogonal acts as a single gateway that lets agents discover, access, and pay for hundreds of APIs without separate credentials or subscriptions. Any agent can instantly discover, access, and pay for hundreds of APIs with no subscriptions, no sales calls, just pay per use. On the supply side, API providers list once and get paid every time an agent calls their API.
The company was founded by Christian Pickett and Bera Sogut, who met at McGill University. Christian worked on payments at Coinbase and billing at Vercel. Bera worked on reCAPTCHA and Maps APIs at Google and is a 2x ACM ICPC World Finalist who also worked at Amazon Robotics.
The product is already live. Developers and agents get instant access to hundreds of APIs through Orthogonal's MCP server or SDK. The platform has 30+ named API partners already integrated, including Exa, CrustData, Apollo.io, and AgentMail. It's MCP-native, meaning it works with Claude, Cursor, and other agent environments out of the box.
"Orthogonal is a perfect example of making something agents want. Agents won't want to sign up for 50 different APIs, managing accounts and keys and rate limits at each one. They'll want to use Orthogonal. So Orthogonal is agent-pilled, but it's also a marketplace and an abstraction layer over all the things agents may want to do on the internet. We were very excited to back these outlier founders."
The underlying bet is that as AI agents proliferate, they'll need a way to transact with the internet's existing services programmatically and on demand. API providers are currently leaving money on the table because billions of agents could be paying for their APIs, but have no way to find or transact with them. Orthogonal positions itself as the connective layer between those two sides: agents that need tools and providers that want distribution.
The company is starting with search, scraping, datasets, and AI model APIs and expanding from there, with a pay-per-request model where every API response includes its price. For teams already building with agents, the pitch is simple: one key replaces dozens.