Reduck
Browser-automation platform enabling AI agents to discover, create, and execute web workflows on sites without APIs.
Website: https://reduck.ai/
Cover Block
From the public record
| Field | Value |
|---|---|
| Name | Reduck |
| Tagline | Browser-automation platform enabling AI agents to discover, create, and execute web workflows on sites without APIs [Tech.eu, October 2026] |
| Headquarters | San Francisco Bay Area |
| Founded | 2025 [Tech.eu, October 2026] |
| Stage | Pre-Seed [Tech.eu, October 2026] |
| Business Model | API / Developer Platform |
| Industry | Other |
| Technology | AI / Machine Learning |
| Geography | Global / Remote-First |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (3+) [Tech.eu, October 2026] |
| Funding Label | Pre-seed [Tech.eu, October 2026] |
| Total Disclosed | ~$3,000,000 [Tech.eu, October 2026] |
Links
From the public record
- Website: https://reduck.ai/
- LinkedIn: https://www.linkedin.com/company/reduck-ai
- Pricing: https://reduck.ai/pricing
The Short Version
PUBLIC Reduck is building a browser-automation layer for AI agents, and it merits investor attention now because it is targeting a persistent integration gap: many business workflows still live on websites that do not expose usable APIs [Tech.eu, October 2026] [LinkedIn, retrieved 2026]. Founded in 2025, the company emerged from a team that had previously worked together at H Company, with Daniel Huynh as CEO, Louis Abraham as CTO, and Raphal Millet as COO according to public company and founder materials [Tech.eu, October 2026] [LinkedIn, June 2026] [LinkedIn, retrieved 2026].
The product, Reduck MCP, is described as a platform that lets an agent explore a website, generate a deterministic browser script, and expose that script as an MCP-compatible tool or endpoint, with the additional claim that execution can happen inside the users own Chrome session rather than through shared credentials [LinkedIn, retrieved 2026] [Station F, September 2026]. That positioning is the clearest point of differentiation in the current materials: Reduck is not presenting itself as a general-purpose agent shell, but as infrastructure for reliable web actions on sites with authentication, cookies, CAPTCHAs, and no native API support [LinkedIn, retrieved 2026].
The early commercial picture is still thin, but directionally coherent. Public claims point to use cases across government portals, e-commerce monitoring, and legacy business software, while the pricing page shows a developer-friendly entry point with a free tier and a $20 monthly plan, suggesting an initial bottoms-up motion before any heavier enterprise packaging [LinkedIn, June 2026] [reduck.ai, retrieved 2026].
Funding is straightforward on the public record: Tech.eu reported a $3 million pre-seed in October 2026 led by Daphni, with participation from Drysdale Ventures and Galion.exe, while Dealroom also lists a 2026 pre-seed of the same amount [Tech.eu, October 2026] [Dealroom.co]. Over the next 12 to 18 months, the key items to watch are whether Reduck can convert product claims into repeatable usage beyond demos, whether deterministic browser scripts hold up across fragile web surfaces, and whether the company can move from low-cost developer adoption to durable paid workflows at scale [Tech.eu, October 2026] [Station F, September 2026].
Single-source, plausible -- Section relies on one independent funding report, partial database corroboration, and several company-controlled or founder-controlled public sources.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Pre-Seed |
| Business Model | API / Developer Platform |
| Industry / Vertical | Other |
| Technology Type | AI / Machine Learning |
| Geography | Global / Remote-First |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (3+) |
| Funding | Pre-seed, total disclosed about $3,000,000 |
The Company in Brief
PUBLIC
Reduck appears to be an early browser-automation infrastructure company formed in 2025, with a public footprint split between a San Francisco Bay Area headquarters reference and a French operating presence through Station F hiring pages [LinkedIn, retrieved 2026] [Station F, September 2026]. The plain-language pitch is consistent across its own materials: the company is building Reduck MCP so AI agents can discover, create, and execute workflows on websites that do not offer usable APIs [reduck.ai, retrieved 2026].
The chronology is still short, but it is readable. Public launch materials and founder profiles place the company in market during 2026, after a period of iteration in stealth described by co-founder Louis Abraham [LinkedIn, June 2026] [LinkedIn, retrieved 2026]. By September 2026, Station F job postings showed the company recruiting in Paris and stated that it had raised more than €2.6 million, and by October 2026 Tech.eu reported a $3 million pre-seed led by Daphni with participation from Drysdale Ventures and Galion.exe [Station F, September 2026] [Tech.eu, October 2026].
A few details remain narrower than an institutional reader would usually want. The available public materials do not establish a legal entity name through a filing source in this dataset, and the Bay Area headquarters reference sits alongside French recruiting evidence rather than replacing it [LinkedIn, retrieved 2026] [Station F, September 2026]. Even so, the operating outline is clear enough for a pre-seed company: founded in 2025, publicly launched in 2026, hiring across product and growth, and positioning itself as infrastructure for agentic web workflows where APIs are missing [reduck.ai, retrieved 2026] [Station F, September 2026].
Single-source, plausible -- Confirmed by company website and Station F, with partial corroboration from LinkedIn and Tech.eu.
What They Have Built
Product surface
MIXED Reduck is selling a plain but consequential promise: let an AI agent operate websites that were never built as APIs, then turn that behavior into a reusable tool [Tech.eu, October 2026] [LinkedIn]. Public product descriptions say Reduck MCP can discover, create, and execute web workflows on sites without APIs, with the workflow exposed as an MCP tool or endpoint for use by applications, cron jobs, or other agents [LinkedIn] [Tech.eu, October 2026].
The public evidence is still mostly company-authored, so the right reading is capability claim rather than field-proven standard. Reduck says scripts can run in a users own Chrome through its extension, preserving logged-in sessions and avoiding direct credential sharing, and it also markets support for authentication, cookies, CAPTCHAs, and sites without APIs [Station F, September 2026] [LinkedIn]. On its website, the company lists compatibility with Claude, Codex, and other MCP-native tools, alongside official scripts for LinkedIn, Twitter, and Reddit, plus workflow examples such as reposting a LinkedIn post from its permalink or extracting followed people and pages from a signed-in LinkedIn account [reduck.ai] [reduck.ai, retrieved 2026].
Performance and packaging
MIXED The core technical angle appears to be determinism rather than open-ended browsing. Tech.eu reports Reduck claims predefined scripts can make agent workflows "three times faster and cheaper" than step-by-step website navigation for tasks such as invoice collection and LinkedIn market research, which fits the broader product positioning around converting exploratory browser behavior into stable scripts [Tech.eu, October 2026]. That matters because browser automation for agents often fails at the point where a demo leaves the happy path and meets login state, brittle page structure, or repeated execution at scale.
Packaging is already visible, even if adoption data is not. Reducks pricing page shows a free tier with 100 script runs per month and a $20 per month plan with 1,000 runs and 10 hours of cloud browser time, suggesting the company is testing both developer self-serve usage and a path toward heavier paid workloads [reduck.ai, retrieved 2026]. The available public record does not support a deeper stack analysis beyond browser extension and cloud execution surfaces, and there is no verified public demo in the source set that would justify stronger claims about reliability, latency, or enterprise controls.
Unconfirmed -- This section relies materially on company website, LinkedIn, and hiring-page product claims, with limited independent corroboration from Tech.eu and Station F.
Market Size and Demand
From the public record The market matters now because general-purpose AI agents are colliding with a stubborn constraint: much of the commercial web still exposes workflows through browser interfaces rather than clean APIs, which keeps browser automation relevant even as model capability improves [Tech.eu, October 2026] [LinkedIn, retrieved 2026].
Reduck sits at the intersection of several adjacent markets rather than inside one neatly reported category. The closest public analogues are robotic process automation, browser automation, developer tooling for integrations, and the newer layer of agent infrastructure that turns software actions into callable tools. The company’s own positioning is narrower than broad RPA, it is focused on letting an agent discover, generate, and run deterministic browser scripts on sites without APIs, then expose those scripts through MCP-compatible tooling [LinkedIn, retrieved 2026] [reduck.ai, retrieved 2026]. That suggests the relevant demand pool is not every automation budget, but the subset of workflows where API access is absent, incomplete, unstable, or too costly to build around [Tech.eu, October 2026] [Station F, September 2026].
There is no cited third-party TAM, SAM, or SOM for Reduck’s exact category in the supplied evidence, so the available view is necessarily analogous rather than direct. Public materials point to early use cases across government portals, legacy business software, e-commerce monitoring, invoice retrieval, and research workflows, all of which are categories where browser interaction is often the system of record even when downstream teams want structured data [LinkedIn, June 2026] [bebee.com, retrieved 2026]. That does not establish category size, but it does identify a broad surface area of repetitive tasks where browser-based execution can substitute for missing integrations.
| Market lens | What the evidence supports | Source quality |
|---|---|---|
| Core market | AI-agent browser automation on websites without APIs | Company and coverage descriptions only [Tech.eu, October 2026] [LinkedIn, retrieved 2026] |
| Adjacent market | Workflow automation across legacy web systems, portals, and vendor sites | Use-case evidence from hiring and founder posts [Station F, September 2026] [LinkedIn, June 2026] |
| Substitute market | Traditional API integrations and broader RPA platforms | Inference from product positioning against sites lacking APIs [Tech.eu, October 2026] [LinkedIn, retrieved 2026] |
| Initial buyer surface | Technical and operational teams automating repetitive browser tasks | Public examples, but no named customers [LinkedIn, June 2026] [bebee.com, retrieved 2026] |
The useful read from the table is that the market case is real but still category-first rather than company-validated. Reduck is addressing a visible integration gap, yet the public evidence does not size that gap with third-party market data or named customer spend.
Demand drivers in the current evidence are practical, not theoretical. First, AI agents need reliable execution layers, and browser navigation remains brittle if every task must be completed step by step in real time; Reduck argues predefined deterministic scripts can make those workflows "three times faster and cheaper" for tasks such as invoice collection and LinkedIn research [Tech.eu, October 2026]. Second, many target systems appear to be exactly the kinds of environments that are hard to modernize, government portals, legacy vertical software, and vendor websites, where teams still need automation regardless of whether the software owner provides an API [LinkedIn, June 2026] [Station F, September 2026]. Third, the company’s Chrome-extension approach is framed as preserving logged-in sessions and avoiding direct credential exposure, which aligns with buyer preference for faster deployment when credentials and session handling are the main friction points [Station F, September 2026].
The adjacent and substitute markets deserve equal attention because they shape both upside and risk. On one side are broad automation tools, integration platforms, and enterprise RPA products that already sell the promise of reducing manual web work. On the other are API-first developer tools and native application connectors that can make browser automation unnecessary when direct access exists. Reduck’s public wedge is the space in between: websites with no API, fragile interfaces, or edge workflows where agent-native browser scripting is simpler than building a custom integration stack [LinkedIn, retrieved 2026] [Tech.eu, October 2026]. If that wedge holds, the company can ride agent adoption without needing to replace the full automation stack.
Macro and regulatory forces are present, though still mostly indirect in the public record. The privacy angle is part of the company narrative, with public references to a privacy-first approach and a product architecture that keeps scripts running in the user’s own browser session rather than routing credentials through a third-party operator layer [AI Time Journal, Mar 19, 2026] [Station F, September 2026]. That could matter in regulated or security-conscious environments, but the current sources stop short of proving compliance adoption. A second force is platform dependence: products that automate third-party websites are exposed to interface changes, rate limits, anti-bot controls, and terms-of-service constraints, particularly when they advertise support for cookies, CAPTCHAs, and social platforms such as LinkedIn, Twitter, and Reddit [LinkedIn, retrieved 2026] [reduck.ai, retrieved 2026]. In practice, that means category demand may expand with agent adoption while operational complexity rises at the same time.
Unconfirmed -- Market framing relies primarily on company materials, founder posts, hiring pages, and one reported funding article; no independent third-party market-sizing report was identified in the supplied sources.
Who Else Is Fighting for This
Market structure
MIXED Reduck is competing less against a single named startup peer in the available record than against three practical alternatives: direct API integrations where APIs exist, human-operated browser workarounds, and broader automation stacks that can already orchestrate tasks but are weaker on API-less web execution [LinkedIn, retrieved 2026] [Tech.eu, October 2026].
That matters because the company’s wedge is narrow but concrete. Public materials consistently position Reduck MCP around websites that do not expose usable APIs, where agents still need to log in, traverse interfaces, and complete repeatable actions in a deterministic way [LinkedIn, retrieved 2026] [Station F, September 2026]. In that lane, the real incumbent is often not another AI-native vendor at all, but internal operations staff clicking through portals, scraping partial data, or handing off brittle scripts between teams.
The adjacent substitute set is easier to describe than the direct venture-backed peer set, because the captured sources do not name competitors. One substitute is conventional API-first integration, which remains superior wherever an official endpoint exists because it is generally cleaner to maintain and easier to govern than browser automation [LinkedIn, retrieved 2026]. Another is generic workflow software and cron-based orchestration, which Reduck itself appears to treat as a complement rather than a head-on target, given its positioning around exposing browser scripts as MCP tools or endpoints for other applications and agents [LinkedIn, retrieved 2026].
Edge and durability
MIXED Reduck’s current edge appears to be product architecture rather than brand, scale, or channel control. The strongest public differentiation is the claim that an agent can explore a site, generate a deterministic browser script, and run that script in the user’s own Chrome session, which preserves logged-in state and reduces the need to hand credentials to a third party [LinkedIn, retrieved 2026] [Station F, September 2026].
If that workflow works reliably, it addresses a real friction point in agent adoption. Many browser tasks fail not because the action is conceptually hard, but because authentication, cookies, CAPTCHAs, and stateful navigation make one-shot autonomous browsing unreliable; Reduck is explicitly marketing support for those constraints rather than treating them as edge cases [LinkedIn, retrieved 2026]. The company also has a pricing surface and a free tier in market, which suggests it is trying to capture developers early instead of waiting for enterprise procurement cycles [Reduck, retrieved 2026].
The durability of that edge is less settled. Browser automation features can be perishable if larger orchestration platforms, model tool ecosystems, or open-source projects absorb the same capabilities, especially when the moat rests on execution quality rather than a proprietary network or exclusive dataset. Reduck’s talent signal is directionally positive, with three co-founders and evidence they previously worked together at H Company, but the public record does not yet show a distribution advantage, regulatory shelter, or installed enterprise footprint that would make displacement especially difficult [LinkedIn, retrieved 2026] [Tech.eu, October 2026].
Exposure and pressure points
MIXED The company is most exposed where the buyer can avoid browser automation altogether. If an official API exists, or if a system of record can be integrated through standard middleware, Reduck’s value proposition narrows because browser-based execution is usually a second-best option chosen for coverage rather than elegance [LinkedIn, retrieved 2026].
It is also exposed to trust and maintenance pressure. The same features that make the product useful, authenticated sessions, logged-in actions, and interaction with sites that lack APIs, also create operational questions around reliability, site changes, terms-of-service sensitivity, and long-term maintenance burden. Public materials frame the approach as deterministic and faster than step-by-step navigation, including a claim of workflows being "three times faster and cheaper" for some tasks, but that claim appears in funding coverage and is not independently benchmarked in the available sources [Tech.eu, October 2026].
A second pressure point is channel ownership. Reduck says it is compatible with Claude, Codex, and other MCP-native environments, which is useful for adoption, but compatibility is not control [Reduck, retrieved 2026]. If model platforms, browser vendors, or orchestration hubs make native authenticated browsing materially easier, Reduck could end up supplying a feature that larger ecosystems subsume.
18-month scenario
MIXED The most plausible 18-month competitive outcome is a sorting of the category into API-first winners on one side and browser-native specialists on the other. In that scenario, direct API integration remains the winner where structured access is available, while Reduck can win a narrower but meaningful segment if enterprises keep encountering high-value workflows inside government portals, legacy software, commerce back offices, and other systems that still do not expose usable endpoints [LinkedIn, June 2026] [Tech.eu, October 2026].
Under that setup, the likely winner if execution quality holds is Reduck itself, specifically if its deterministic-script approach proves more reliable than free-form agent browsing across repeated tasks. The likely loser if official APIs continue to proliferate, or if orchestration platforms fold similar browser tooling into broader suites, is the stand-alone browser layer that cannot show lower failure rates or lower maintenance than adjacent tools. The evidence base here is thin on named rivals, so the competitive view is clearer on category pressure than on company-by-company share shifts.
Unconfirmed -- This section relies mainly on company materials, LinkedIn posts, pricing pages, a Station F job listing, and one independent funding report; no named direct competitors were confirmed in the provided sources.
Opportunity
PUBLIC
The prize here is larger than a workflow tool: if Reduck can become the reliable interface layer between AI agents and the long tail of websites that still do not expose usable APIs, it has a path to matter anywhere software still depends on clicking through a browser [Tech.eu, October 2026] [LinkedIn, retrieved 2026].
The headline opportunity is straightforward. Reduck is trying to turn the browser itself into programmable infrastructure for agents, with deterministic scripts that can be exposed as MCP tools or endpoints rather than forcing an agent to improvise each step on the fly [LinkedIn, retrieved 2026]. That matters because a meaningful share of business workflows still sit inside government portals, legacy business software, e-commerce dashboards, and other web interfaces that are accessible to humans but awkward for software, and Reduck is explicitly positioning around that gap with support for authentication, cookies, CAPTCHAs, and sites without APIs [LinkedIn, retrieved 2026] [Station F, September 2026]. The public evidence is still early and largely company-sourced, but it is enough to make the upside legible: Reduck does not need to replace existing APIs, only to become the default option when no clean API exists and the browser is the only practical surface [Tech.eu, October 2026] [LinkedIn, June 2026].
The plausible scale paths are narrower than broad horizontal automation at first, but they are real if product reliability holds.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Default browser layer for agent builders | Developers adopt Reduck as the standard way to connect agents to websites without APIs, making it a recurring dependency inside agent stacks | MCP adoption broadens across tools and Reduck's existing compatibility with Claude, Codex, and other MCP-native environments translates into embedded usage [Reduck, retrieved 2026] | Reduck already markets itself as an MCP browser-automation platform and exposes scripts as tools or endpoints rather than one-off automations [LinkedIn, retrieved 2026] [Tech.eu, October 2026] |
| Wedge through hard-to-automate vertical workflows | The company wins dense use cases in government, compliance, procurement, and legacy software, then expands account scope from one script to many | A few repeatable templates for tax portals, invoice collection, procurement monitoring, or industry-specific back offices prove materially faster than manual browser navigation [LinkedIn, June 2026] [Station F, September 2026] [Tech.eu, October 2026] | Public examples already point to tax-message retrieval, e-commerce price monitoring, legacy dental portals, and invoice collection, which suggests the wedge is not theoretical [LinkedIn, June 2026] [Station F, September 2026] |
| Script marketplace and distribution layer | Reduck evolves from a tool vendor into a distribution surface where teams reuse official and community scripts across common sites | Growth in official script coverage for properties such as LinkedIn, Twitter, and Reddit, combined with low-friction pricing and free-tier adoption, drives reuse [Reduck, retrieved 2026] | The company already lists official scripts, a free tier with 100 runs a month, and a paid self-serve plan, which are the early ingredients for bottom-up adoption and packaged reuse [Reduck, retrieved 2026] |
What compounding could look like is not classic social network effect so much as workflow density. Each successful script reduces the cost of serving the next adjacent use case on the same site or class of sites, and each additional supported destination makes the platform more useful to agent builders who need reliability over improvisation [LinkedIn, retrieved 2026] [Tech.eu, October 2026]. If the extension-based architecture continues to preserve logged-in sessions in the user's own Chrome, that may also lower one of the main adoption frictions in browser automation, namely credential handling and session breakage [Station F, September 2026]. The public signs of an early flywheel are modest but visible: Reduck has launched self-serve pricing, highlighted predefined scripts, and surfaced use cases across several workflow categories rather than a single narrow demo [Reduck, retrieved 2026] [LinkedIn, June 2026].
The size of the win is difficult to benchmark precisely from public data because no market-sizing source or direct public comparable is confirmed in the record provided here. Still, a reasonable framing is that infrastructure companies that become the default connective tissue for developers can support venture-scale outcomes even before they own the full application layer, and Reduck's $3 million pre-seed led by Daphni indicates outside investors see at least that possibility at a formative stage [Tech.eu, October 2026] [Dealroom.co]. If the first scenario plays out and Reduck becomes a standard browser-access layer for agentic workflows, the upside could be a platform-scale infrastructure business rather than a services tool, which would support a venture outcome in the hundreds of millions of dollars or more in enterprise value (scenario, not a forecast). That inference rests on business model shape and product position, not on disclosed revenue, valuation, or market share, none of which are established in public sources here.
Unconfirmed -- This section relies materially on company statements, LinkedIn posts, Station F hiring materials, and one main funding report from Tech.eu, with no independently confirmed market sizing or operating metrics in the public record used here.
Sources
From the public record
[Tech.eu, October 2026] Reduck secures $3M to make web automation faster for AI agents | https://tech.eu/2026/10/22/reduck-secures-3m-to-make-web-automation-faster-for-ai-agents
[LinkedIn, retrieved 2026] Reduck AI LinkedIn Company Page | https://www.linkedin.com/company/reduck-ai
[LinkedIn, June 2026] Louis Abraham LinkedIn Post | https://www.linkedin.com/posts/louis-abraham_for-months-we-iterated-in-stealth-we-kept-activity-7467100161110917120-xrg3
[Station F, September 2026] Vibe Coder Intern - Paris | https://jobs.stationf.co/companies/conception-dev/jobs/stage-vibe-coder_paris
[reduck.ai, retrieved 2026] Make your agent browse the web fast and reliably | Reduck | https://reduck.ai/
[Dealroom.co] Reduck | https://dealroom.co/companies/reduck
[LinkedIn] Reduck AI LinkedIn Company Page | https://www.linkedin.com/company/reduck-ai
[reduck.ai] Make your agent browse the web fast and reliably | Reduck | https://reduck.ai/
[bebee.com, retrieved 2026] Chief of Staff Intern at Reduck (Station F, Paris, Ile de) | https://bebee.com/fr/jobs/chief-of-staff-intern-at-reduck-station-f-paris-ile-de--t7xk-790206801
[AI Time Journal, Mar 19, 2026] Building and Leading a Privacy-First Startup - Interview with Daniel Huynh · AI Time Journal | https://aitimejournal.com/interview-with-daniel-huynh-ceo-mithril-security
[Reduck, retrieved 2026] Pricing: free, $20/month, or custom | Reduck | https://reduck.ai/pricing
Articles about Reduck
- Reduck’s Browser Extension Turns the Web Into an Agent Playground — The French-American startup’s $3 million round, led by Daphni, funds a bet that automation will live in the browser tab, not the cloud.