Video tutorial · The Metaverse Guy
TinyFish: The AI Platform Replacing Playwright, Firecrawl & BrowserBase
I test TinyFish, an agentic web infrastructure platform that combines search, fetch, browser and agent under one API for AI-driven scraping.
This video was made in partnership with TinyFish. It carries YouTube's paid-promotion label.
- Published
- 27 May 2026
- Length
- 7 min
- Views
- 10,851 views
In this video, I test TinyFish, a new platform that aims to replace multiple scraping tools. I build an AI competitor tracking pipeline that logs into a SaaS portal, navigates a JavaScript-heavy dashboard, extracts pricing data, and returns structured JSON. The platform also integrates with AI coding agents like Cursor and Claude Code.
Why Traditional Web Scraping Breaks
If you build AI agents or automate browser workflows, you know how fragile traditional scraping can be. A Playwright script works until a CSS selector changes, Cloudflare blocks you, or sessions break. Suddenly your entire workflow stops working.
Stitching together separate tools for search, rendering, extraction, browser automation, and anti-bot handling adds complexity and maintenance overhead. TinyFish aims to solve this by handling the full workflow in one place.
What is TinyFish?
TinyFish is an agentic web infrastructure platform that combines search, fetch, browser, and agent capabilities under a single API. Instead of juggling multiple services, you get a unified interface for finding pages, rendering JavaScript, extracting data, and managing authenticated sessions.
It’s designed to work seamlessly with AI coding agents like Cursor and Claude Code. The TinyFish CLI can be integrated directly into your workflow using a SKILL.md file, giving your local AI agent access to the full API surface without extra SDK setup.
Building an AI Competitor Tracking Pipeline
In the demo, I built a full AI competitor tracking pipeline. The workflow found a SaaS pricing portal, logged in through an authenticated browser session, navigated a JavaScript-heavy dashboard, extracted pricing data, and returned structured JSON.
The platform handled the anti-bot challenges and session management automatically, so I didn’t need to write custom logic for Cloudflare or CAPTCHAs. The output was clean, structured data ready for downstream use.
Using TinyFish with AI Coding Agents
One of the most impressive parts is how fast TinyFish works with AI coding agents. I used the TinyFish CLI with Cursor and Claude Code. By adding a SKILL.md file to the project, the AI agent could directly call the TinyFish API without any additional SDK configuration.
This means you can instruct your coding agent to perform complex web automation tasks in natural language, and TinyFish executes them reliably.
Scaling and Final Thoughts
The pipeline scaled to 50 concurrent browser sessions without issues. For anyone building AI agents, browser automation workflows, or autonomous AI infrastructure, this level of concurrency and reliability is essential.
TinyFish replaces the need for Playwright, Firecrawl, and BrowserBase in many scenarios by providing a single, agentic platform. If you’re tired of maintaining brittle scraping scripts, it’s definitely worth checking out.
Tools used in this video
- TinyFish
- Playwright
- Firecrawl
- BrowserBase
- Cursor
- Claude Code
Questions
- What does TinyFish replace?
- TinyFish combines the functionality of Playwright, Firecrawl, and BrowserBase into one platform, handling search, fetch, browser automation, and anti-bot measures under a single API.
- Can TinyFish handle authenticated sessions?
- Yes, the demo showed logging into a SaaS portal through an authenticated browser session, navigating a JavaScript-heavy dashboard, and extracting data behind a login.
- How does TinyFish work with AI coding agents?
- The TinyFish CLI integrates with Cursor and Claude Code via a SKILL.md file, giving your local AI coding agent direct access to the full API surface without extra SDK setup.
Topics
Chapters
- 00:00 Intro
- 00:18 Why Traditional Web Scraping Breaks
- 00:48 The Problem with AI Web Automation Today
- 01:18 What is TinyFish?
- 01:52 Building an AI Competitor Tracking Pipeline
- 02:35 Installing TinyFish CLI
- 03:05 Using TinyFish with Cursor & Claude Code
- 03:42 Running the Browser Automation Workflow
- 04:25 Structured JSON Output Demo
- 05:05 Scaling to 50 Concurrent Sessions
- 05:48 Nemo Codification Explained
- 06:28 Why TinyFish Matters for AI Agents
- 07:02 Final Thoughts
- 07:20 Outro
Video description and links
This article was written from the video's description with the help of AI. Watch the video for the full walkthrough.



