browse.sh
A starter page for a fast, scriptable browsing workspace.
AI Web Agent Infrastructure
Summary
TinyFish is a serverless enterprise infrastructure platform providing APIs for AI web agents to search, fetch, navigate, and automate web tasks at scale. It handles authentication, anti-bot bypass, and structured data extraction with a claimed 89.9% accuracy on the Mind2Web benchmark.
Highlights
- Four core APIs: Search (live JSON results), Fetch (rendered page to markdown/JSON), Browser (authenticated stealth navigation), and Agent (end-to-end task automation)
- Claims 89.9% accuracy on Mind2Web benchmark, the highest publicly benchmarked web agent
- Serverless architecture handles hundreds of websites simultaneously with a single API key
- Free API access for agents; used in production by teams at DoorDash, Grubhub, and others
- Handles compound tasks across dynamic sites including form-filling, login walls, and anti-bot protections
auto-generated
Context
Audience
Developers and engineering teams building AI agents, automation pipelines, or data extraction workflows that require reliable web interaction at scale
DomainWeb Automation & AI Agents
FormatSaaS platform with API
AccessFree API access for agents; enterprise tier available
Source Overview
TinyFish offers enterprise infrastructure for AI web agents via a serverless architecture.
It provides APIs for Search (real-time JSON from live web), Fetch (rendered page data in markdown/JSON/HTML), Browser (authenticated navigation bypassing anti-bot measures), and Agent (89.9% accurate task automation like form-filling and extraction). The platform handles compound tasks across dynamic sites with one API key, used by teams at DoorDash, Grubhub, and others for production workflows.
Topics
Related
Mind2Web benchmarkBrowser automationAnti-bot evasionServerless architectureStructured data extractionWeb scraping APIs
