TinyFish is web infrastructure for AI agents that searches, fetches, browses, and completes tasks on live websites. It helps developers and enterprise teams gather current web data, use authenticated portals, and automate browser-based workflows.
TinyFish is a web infrastructure platform for AI agents. It combines four capabilities behind one API key and Wallet: TinySearch for fresh, browser-rendered search results; TinyFetch for converting rendered pages into Markdown, JSON, or HTML; TinyBrowser for driving a cloud browser with Playwright, Puppeteer, or CDP; and TinyAgent for navigating websites, filling forms, authenticating, and returning structured results from a plain-language goal.
The platform is designed for work that depends on the live web rather than a fixed API or cached index. Typical workflows include monitoring pages and prices, checking and enriching records against current sources, researching questions across multiple sites, and completing forms or retrieving files from authenticated portals. TinySearch and TinyFetch are free within their stated limits. TinyAgent and TinyBrowser use prepaid Wallet funds based on usage, with enterprise options for workload-specific limits and controls.
TinySearch queries Google’s live index through a real browser and returns structured JSON rather than an unstructured text result. The pricing page lists limits of 30 requests per minute and 500 requests per hour at no charge.
TinyFetch turns a rendered URL into clean Markdown, JSON, or HTML in one call. It is intended for extracting usable page content without passing navigation, cookie banners, and advertising through to downstream model calls.
TinyBrowser provides a real cloud browser that developers can control with Playwright, Puppeteer, or CDP. It supports persistent state and is priced by minute; the starting account limit is five concurrent sessions.
TinyAgent accepts a goal in plain English, then navigates live sites, fills forms, handles logged-in sessions, and returns structured results. Agent runs include model inference, residential proxy routing, infrastructure-level anti-bot handling, and stored screenshots, logs, and source URLs.
TinyVault lets teams select which credentials a run can use without exposing passwords to the model, while TinyProfiles saves signed-in sessions for later runs. Run recordings provide screenshots and replay links for auditing actions.
TinyFish can be used through an API, SDKs, CLI, and MCP. The site lists integrations for tools and assistants including ChatGPT, Claude, Grok, Cursor, OpenClaw, Dify, n8n, Vercel, AG2, Qoder, and omp, plus an official Python SDK.
Track changes to competitor pages, prices, filings, job posts, suppliers, or regulatory information and generate alerts from current page content.
Check CRM or other internal records against live company and contact sources, then fill gaps or identify incorrect values.
Ask questions that require information from many websites, such as identifying supplier ownership changes, and return answers with links to the sources used.
Sign in to supplier, carrier, payer, or other business portals, complete multi-step forms, and retrieve files without having staff repeat the process manually.
Yes. TinyVault supplies selected credentials to a run while keeping them out of the model’s prompt, and TinyProfiles can preserve a signed-in session. The enterprise page says agents can handle SSO and MFA without reading the password.
Search and Fetch continue to work within their limits because they do not draw from the Wallet. An Agent run already in progress finishes, while new Agent and Browser runs require a deposit.
Search and Fetch are free within their listed limits. Agent costs $0.016 per step and Browser costs $0.002 per minute. There is no subscription or monthly minimum; Agent and Browser usage is deducted from prepaid Wallet funds.
Yes. Accounts start with two concurrent Agent runs and five concurrent Browser sessions. TinyFish says higher concurrency and rate limits can be set for a team’s workload through an enterprise conversation.
Teams can use the REST API, SDKs, CLI, or MCP, or install listed integrations for AI assistants, developer tools, and automation platforms. The integrations page also supports requesting an integration or building one with the REST API.
流量数据仅供参考。
www.simular.ai
Simular is an autonomous computer platform that uses AI agents to operate real computers and automate digital workflows. It serves individuals, builders, operators, and businesses that need to automate tasks across desktop applications and websites, including workflows that do not expose traditional APIs.
www.stagehand.dev
Stagehand is an SDK for building browser agents with Playwright-style APIs, natural-language actions, structured extraction, and self-healing behavior. It supports TypeScript, Python, and Go for agents that need to operate web pages with browser-level control.
anysite.io
Anysite provides live web data to AI assistants and software through MCP, REST API, and CLI interfaces. It helps GTM, marketing, research, and other teams query 900+ sources and 5,000+ endpoints for structured results.
steel.dev
Steel is open-source browser infrastructure for AI agents and web automation. It provides cloud browser sessions through an API, with support for common browser automation tools and workflows such as scraping, research, RPA, and agent-driven web tasks.
developers.cloudflare.com
Kitesurf is Cloudflare’s stateless browser engine for AI agents and bursty automation. It renders web pages and supports actions such as HTML extraction, screenshots, and PDFs while using less CPU and memory than Chromium for common agentic workloads.
www.hyperbrowser.ai
Hyperbrowser provides on-demand cloud browsers and sandboxes for AI agents, web automation, scraping, testing, and agentic code execution. It connects browser workflows through an API, Playwright, Puppeteer, Python, or Node.js.
