Deals · AI infrastructure
Reduck raises $3M pre-seed to give AI agents scripts for the open web
A Paris-based browser automation platform that gives AI agents pre-built scripts to interact with websites and web applications without APIs — reducing cost and failure rates for AI-native workflows that depend on the open web.
Reduck (Paris) has raised $3 million (approximately €2.7 million) in a Pre-Seed round led by Daphni, with participation from Drysdale Ventures and Galion.exe. The company was founded by Daniel Huynh (CEO), Louis Abraham (CTO), and Raphaël Millet (COO) — a trio whose prior track records overlap. Huynh and Millet co-founded Mithril Security, which was acquired by H Company (the French AI lab that has raised over $220 million); Abraham was a core researcher at H Company and an ICPC finalist with prior work in biomedical ML at Stanford. Huynh also created LaVague, an open-source AI web agent. Reduck is their bet that the infrastructure layer for the next generation of those agents is still missing.
1,200 scripts, 250 websites
The problem Reduck is solving is familiar to anyone building AI agents that need to act on the web: browser automation without predefined paths fails at a rate that compounds quickly — agents hallucinate actions, retry on errors, and burn token budgets on interactions a human would handle in seconds. Reduck's platform, Reduck MCP, short-circuits this with a library of 1,200+ ready-to-run scripts across 250+ websites and web applications. When an AI agent needs to interact with a supported site, it uses a tested, predefined path instead of navigating live. The company claims this makes interactions three times faster and cheaper than unguided browser automation.
The platform connects natively to MCP (Model Context Protocol), which means it is accessible to the major AI assistants without bespoke integration. The next phase adds autonomous script generation — the system learns new site patterns as agents encounter them — alongside a dedicated developer offering targeting portals in insurance, healthcare, and internal CRMs where no vendor API will be published.
The H Company alumni
The founder combination is the strongest signal in the round. Huynh, Abraham and Millet have spent the pre-2026 years inside the AI infrastructure buildout — a frontier lab, an open-source agent framework, applied ML research — rather than arriving from outside the problem. Daphni's lead is consistent with the Paris fund's track record in AI-adjacent infrastructure bets. Galion.exe, the French tech-founder network turned VC, adds ecosystem weight: founders backing other founders with demonstrated exit track records.
The $3 million funds a team expansion to eight, a larger script library, the developer-offering rollout, and early work on GEO (generative engine optimisation) automation. Pre-seed stage means Reduck is still proving demand at scale. The H Company alumni relationships provide a warm path to enterprise design partners; the question is whether the use-case concentration in portals that resist MCP adoption is large and durable enough to build a company on.
Sources
Threaded to this story
AI infrastructure ·
Weflayr raises $2.5M to route enterprise AI requests to the right model at the right cost
€2.2M · Pre-Seed
AI infrastructure ·
ZML raises €17.5M to run LLMs on any chip — and break Nvidia's inference grip
€17.5M
Climate / Energy ·
Epyr raises €15M Series A to electrify European industrial heat at scale
€15M · Series A
Every European round, in your inbox by 8am.
The day's seed and Series A rounds across France and Europe — threaded, sourced, and read in two minutes. Free.