TinyFish unifies web primitives
TinyFish unveiled an API that unifies web primitives—search, fetch, browser automation and agent access—to provide live web connectivity for consumer automations. The announcement included a 500‑credit free tier to let developers test productivity workflows (x.com).
Software that uses the live web usually stitches together search, page fetching, and browser control from different vendors. TinyFish is packaging those pieces into one API with one key and one pricing system. (docs.tinyfish.ai) TinyFish’s documentation now lists four public products: Agent, Search, Fetch, and Browser. The company says developers can use the Agent tool for goal-based automation, Search for ranked results, Fetch for rendered page content, and Browser for direct remote browser sessions. (docs.tinyfish.ai) The company’s pricing page says new users get 500 free steps with no credit card required. Its MCP integration page breaks usage into credits, including 1 credit for 2 searches, 1 credit for 15 fetches, 1 credit per run step for automation, and 1 credit for 4 browser-minutes. (tinyfish.ai, docs.tinyfish.ai) A web automation stack is the plumbing behind bots that read pages, click buttons, fill forms, and pull data from sites that do not offer clean application programming interfaces. TinyFish is pitching its stack as a way to avoid managing separate browsers, proxies, and extraction tools for those jobs. (tinyfish.ai, docs.tinyfish.ai) That pitch lands as coding assistants are being pushed beyond text generation into tasks on live websites. TinyFish’s MCP page says Claude, Cursor, and other compatible assistants can connect to its tools so an assistant can search, browse, and extract data on a user’s behalf. (docs.tinyfish.ai) The technical split is straightforward. Search returns links and snippets, Fetch renders a page and extracts text, Browser opens a controllable Chrome session over the Chrome DevTools Protocol, and Agent executes a natural-language task on a real site. (docs.tinyfish.ai, docs.tinyfish.ai, docs.tinyfish.ai, docs.tinyfish.ai) TinyFish is also framing the product around “live” data rather than cached search results. Its website highlights use cases such as checking health-plan portals, monitoring state insurance filings, tracking e-commerce pricing, and aggregating shipment data across carrier sites. (tinyfish.ai, tinyfish.ai) The trade-off is that live browser work is slower and more failure-prone than calling a normal software API, especially on protected or heavily scripted sites. TinyFish’s own docs include anti-bot guidance, browser profiles, and run-status errors such as site blocks, timeouts, and max-step limits. (docs.tinyfish.ai, docs.tinyfish.ai) For developers, the immediate question is whether one vendor can replace a patchwork of search, scraping, and automation tools without raising costs or lowering reliability. TinyFish’s answer is the free starter allotment: enough credits to test whether a single web stack can handle real workflows before paying. (tinyfish.ai)
