French-American AI startup Reduck has raised $3 million in a pre-seed funding round led by Daphni, with participation from Drysdale Ventures and Galion.exe. The company is developing technology that enables AI agents to interact with websites and automate tasks without relying on traditional APIs.

Founded by Daniel Huynh (CEO), Louis Abraham (CTO) and Raphaël Millet (COO), Reduck aims to address one of the challenges limiting the adoption of AI agents: the lack of direct access to websites and business applications.

While AI agents are increasingly being developed to automate digital tasks, many websites and internal business tools do not provide APIs or Model Context Protocol (MCP) interfaces that allow agents to interact with them directly. As a result, agents often rely on browser-based interactions, such as clicking and typing, which can be slow, costly and unreliable when performing complex tasks.

To address this challenge, Reduck has developed Reduck MCP, a platform that allows AI agents to discover, execute and create browser automation scripts. Instead of navigating websites step by step, agents can use predefined scripts to complete tasks, reducing the time and computational resources required.

According to the company, its technology can make agents three times faster and cheaper across tasks such as downloading invoices from multiple vendors and conducting market research on LinkedIn. Reduck's library currently includes more than 1,200 ready-to-run automation scripts covering over 250 websites.

To expand and improve this library, Reduck is developing a system of AI agents capable of creating and maintaining automation scripts autonomously, with a focus on reliability and scalability.

Alongside these developments, the company is preparing a developer offering that will enable teams to build custom AI agent systems connected to applications without APIs, including specialised portals in insurance and healthcare and internal customer relationship management systems. Reduck is also exploring applications for automating knowledge work, starting with generative engine optimisation (GEO).

The new funding has supported the expansion of Reduck's team across AI engineering, full-stack development and infrastructure. The company plans to further improve its automation technology, expand its script library and introduce a new offering for developers.