Indian American co-founders Kaushik Mahorker and Yagnya Patel have launched their startup Wildcard, a gateway for AI agents to use APIs, on Thursday.
Backed by Y Combinator, Wildcard enables agents to select the right application program interface (API) flows for their tasks, with API providers constructing “agents.json” files to help AI agents discover their APIs.
READ: Pulse plans to test and integrate Gemini 2.0 for document processing (February 7, 2025)
Wildcard understands that API infrastructure over the past 20 years or so has been built for developers alone. New applications are agentic and traffic to APIs will come from AI agents. Agents need interfaces that deliver outcomes, yet APIs are limited to discrete actions. Developers end up having to prompt tune, build tool definitions, and chain LLMs to get simple integrations to work. This is where Wildcard comes in.
Wildcard acts as a single tool to find the right API flow to call. With Wildcard’s open source software development kit (SDK), the user can run actions on their own infrastructure wherever their agent lives. The startup is launching 10 curated APIs, including officially verified ones like Resend, Alpaca and more.
Patel took to LinkedIn to share: “Wildcard is translating APIs for agents to use – with agents.json. And now you can try it out!”
Wildcard allows users to start with their collection of “agents.json” specifications, ready to use with AI agents. Users can also create custom collections by adding their own APIs and tools with “agents.json”.
READ: Inside ClickMe: A conversation with Professor Thomas Serre and Jay Gopal on advancing explainable AI (February 19, 2025)
The newly launched software also helps with tool selection with AI agents being able to query it to pick tools using natural language. Wildcard also drops in tool definitions catered to your agent’s model. It allows users to get their own tool calls, and execute actions on their own infrastructure.
Mahorker and Patel had first met as undergrads. While Mahorker is a former engineering manager at Scale AI where he led the GenAI Allocation team, Patel previously worked on NLP and Knowledge Graphs at Tesla, Amazon, and Truveta.