How TinyFish launched OpenSource Free WebAccess on X

TinyFish Search and Fetch now power the free first call in WebAccess, an open-source path Jason Zhou shipped through treg.to for people building agents. An agent can search the live web and read a page with no credit card. TinyFish stated free search at 30 requests a minute for this integration. Fetch renders the page in a real browser, including JavaScript-heavy sites, and returns markdown, JSON, or HTML with navigation, scripts, and banners stripped out, so the model is not handed a full page of chrome. If that free call comes back empty or blocked, WebAccess tries the next vendor for the same job, ordered by hit rate, price, and speed, and stops at the first real answer. A vendor that charges only on success costs nothing when it misses. The response includes a trace of who was tried, who answered, and what the call cost. The same API can also return structured fields from platforms such as X, LinkedIn, Reddit, YouTube, and Amazon, so the agent does not have to pull those sites as raw HTML. Callers can prefer TinyFish, skip a vendor, cap fallback spend, or stop after the first miss using request headers. The product page prices 10,000 calls, split as 5,000 searches and 5,000 page reads, at $0 on the free path, against $40 on Exa, $75 on Firecrawl, and $96 on Tavily. The web-access skill and source are in the treg repository on GitHub.