Last updated: 2026-05-16
Discover TinyFish, an AI workflow automation platform. Features, pricing, and alternatives on hokai.io.
NotebookLM (Google) is a free AI research assistant for analyzing documents and generating insights. Offers audio overviews of source materials with no subscription required.
TinyFish is enterprise web agent infrastructure launched from stealth in August 2025 with $47 million Series A funding. Founded by Sudheesh Nair (former Nutanix President), Shuhao Zhang (ex-Meta engineering leader), and Keith Zhai (ex-Wall Street Journal), TinyFish enables organizations to automate complex web interactions at scale without managing browsers, proxies, or infrastructure. The platform operates as a serverless API that accepts natural language goals and URLs, then deploys AI agents to navigate websites, handle authentication, extract structured data, and complete transactions—all in real-time with live streaming progress. Unlike traditional web scraping or browser automation tools that require CSS selectors, XPath, or complex configuration, TinyFish agents work from plain English descriptions of tasks. A user describes "find the cheapest hotel in Tokyo" and the agent navigates, compares, and returns structured JSON without code. TinyFish's core infrastructure delivers parallel execution across up to 1,000 simultaneous operations, authentication handling through logins, forms, and paywalls, live data extraction from dynamic sites (JavaScript-rendered content included), anti-bot protection and stealth modes for harder sites, and cost-all-inclusive pricing with no separate charges for browsers, proxies, or AI inference. The platform scores 82 percent on the Mind2Web benchmark—a 39-point lead over OpenAI's Operator (43%) and demonstrates production readiness. The platform integrates natively with Claude via Model Context Protocol (MCP), enabling seamless connection to Claude Desktop. It also provides REST API, Python and JavaScript SDKs, and integrations with n8n, AG2, and workflow automation tools. AgentQL, the platform's query language, allows developers to extract data and automate workflows with precision, supported by a browser debugger and recipe cookbook. TinyFish is already in production at Fortune 500 companies running 30+ million operations monthly. Google uses TinyFish agents to aggregate hotel inventory from thousands of properties in Japan for Google Hotels search—data previously inaccessible due to underlying infrastructure limitations. DoorDash manages web interaction complexity at scale. ClassPass, Amazon, Amplemarket, NextEra, Jobright, and Healdev deploy TinyFish for healthcare (prior authorization tracking, clinical trial data), insurance (regulatory filing monitoring), financial services, retail (price monitoring, MAP violation detection), and supply chain (shipment tracking). Pricing is transparent and includes all infrastructure: free tier with 500 steps, pay-as-you-go at $0.015 per step with 2 concurrent agents, Starter plan at $15/month (1,650 steps, 10 agents), Pro plan at $150/month (16,500 steps, 50 agents), and Enterprise with custom volume and support. Overage costs decrease with volume: $0.014 per step on Starter, $0.012 per step on Pro. All plans include 30-day run history with observability and screenshots, priority support, and Workbench, API, and MCP access.
Free: 500 steps, no credit card required. Pay-as-you-go: $0.015/step, 2 concurrent agents. Starter: $15/month (1,650 steps, 10 concurrent agents, $0.014 overage/step). Pro: $150/month (16,500 steps, 50 concurrent agents, $0.012 overage/step). Enterprise: Custom pricing with custom volume, concurrency, and SLA. All-inclusive: no separate charges for browsers, proxies, or LLM inference.
TinyFish is an enterprise web agent platform that automates business workflows across websites at scale. Founded in 2023 by Sudheesh Nair (ex-Nutanix), Shuhao Zhang (ex-Meta), and Keith Zhai (ex-WSJ), TinyFish launched from stealth in August 2025 with $47M Series A funding. Users describe tasks in plain English and TinyFish deploys AI agents to navigate sites, handle authentication, extract data, and complete transactions—all without managing browsers, proxies, or infrastructure. The platform runs 30M+ operations monthly in production for Google Hotels, DoorDash, ClassPass, and other Fortune 500 companies.
TinyFish offers transparent step-based pricing with no hidden charges for browsers, proxies, or LLM inference. Free tier: 500 steps, no credit card required. Pay-as-you-go: $0.015/step with 2 concurrent agents. Starter: $15/month (1,650 steps, 10 concurrent agents, $0.014 overage/step). Pro: $150/month (16,500 steps, 50 concurrent agents, $0.012 overage/step). Enterprise: Custom pricing with custom volume, concurrency, and SLA. All-inclusive pricing means no separate charges for browsers, proxies, residential IPs, or AI inference.
Core features include parallel execution across up to 1,000 simultaneous operations, natural language task control (no CSS selectors or XPath), authentication handling (logins, forms, paywalls), live data extraction from dynamic JavaScript-rendered sites, anti-bot protection and stealth modes, and real-time progress streaming via Server-Sent Events. Integration with Claude Desktop via Model Context Protocol (MCP) enables direct AI assistant access. AgentQL query language, Python and JavaScript SDKs, REST API, n8n integration, and Playwright support cover developer needs. 30-day run history with observability and screenshots support debugging.
Yes, TinyFish offers a free tier with 500 steps and no credit card required. The free tier includes full access to core features (parallel operations, authentication handling, live data extraction) and integrations, making it suitable for testing and small workflows. Once you exceed 500 steps monthly, you can switch to pay-as-you-go ($0.015/step) or subscribe to Starter ($15/month) or higher tiers.
Competitors include CloudCruise, Nebula, Zyte (traditional web scraping), Airtop (cloud-hosted browsers), Hexomatic (web data extraction), Axiom AI (Chrome extension automation), OutWit (data extraction), and Diffbot (AI-powered web crawling). Airtop is simpler for no-code users. Hexomatic and Axiom AI are cheaper for single-site automation. TinyFish excels at enterprise-scale parallel automation (1,000+ concurrent operations) with native AI integration and all-inclusive pricing, making it strongest for Fortune 500 workflows.
TinyFish is best for Fortune 500 companies automating complex web workflows at scale, healthcare organizations tracking prior authorizations across provider portals, insurance firms monitoring regulatory filings, financial services extracting data from banking platforms, and retail companies monitoring prices and MAP violations. It is less suitable for freelancers, small teams, or sites with heavy CAPTCHA protection, as CAPTCHA solving is not currently supported.
Yes, TinyFish offers a comprehensive API with REST endpoints, Python and JavaScript SDKs, and integrations with n8n, AG2, and Playwright. The platform also supports Claude Desktop via Model Context Protocol (MCP) for direct AI assistant integration. API documentation is available at https://docs.tinyfish.ai, and open-source recipes and examples are published on GitHub at https://github.com/tinyfish-io/tinyfish-cookbook.