Changelog

SiteGround fronts a large chunk of small WordPress and WooCommerce shops, so its Robot Challenge Screen is a class of targets, not one site. With unblocker: true, Single now solves it in-process and returns the real page. Before this, you had to bring up Browser for that class of target, which meant paying for a browser session and the wait. The clearance cookie comes back in the response, so a replay stays cheap.

SiteGround puts a 'Robot Challenge Screen' in front of a lot of small WordPress and WooCommerce shops; there's nothing to click, it's a SHA-1 proof of work. Single now clears it in one call and returns the real page, at 2 credits instead of a 30-credit browser session. It's the first non-Cloudflare defense Single handles on its own.

Every plan limit now sits next to what you've actually used this billing period. Credits split into billed (successful requests only) and spent (everything you consumed), with a per-key breakdown and live counters for concurrency and rate. Open it from the sidebar in your Dashboard.

The Dashboard has a new Billing details tab for your standing invoice profile: company, VAT with live verification, and address. Organizations now lives on its own page as a sortable table you can create from and rename. Every API key has an owner (personal or an organization), and you can transfer a key between them without losing its metrics.

If we've built you a custom plan, it now shows up as its own tab when you change plans. And if you don't have a subscription yet, that same picker takes you straight to checkout. No more email-and-wait to move onto your plan.

FourA MCP 0.5.0 adds strict country-scoped selection to foura_proxy. Pass exitCountries with the target-visible countries you need, and a scoped success returns exitCountry plus the reusable proxy ID for foura_browser. If the pool has no match, no_eligible_proxy preserves your scope instead of falling back to another country.

Pass exitCountries: ["CZ", "GB"] to /api/proxy to keep the exit inside the list you provide, and the response reports which country actually served the request so follow-up calls can pin the same geography. If the pool has no match, the call fails with no_eligible_proxy instead of falling back to a different country. Try it without any code in the Playground.

The Overview page now shows a Credits panel with your total credits used plus a per-product breakdown (Single, Proxy, Browser). Big numbers stay compact (like 29.7M) and hovering reveals the exact value. Respects your active filters and time range.

There's now a dedicated home for the FourA MCP Server at foura.ai/mcp. Pick your client (Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, or the hosted remote server), copy the config, or install with one click through the Add-to-Cursor and Install-in-VS-Code deeplinks. Four tools ship with it: foura_auto, foura_single, foura_proxy, and foura_browser.

Some targets only respond to specific countries (regional sportsbooks, marketplaces with geo restrictions). Auto now notices when a site is geo-gated, remembers which countries return real responses, and pulls fresh exits from those countries on later calls. Concurrent load also spreads across the allowed set, so one exit doesn't take the whole load and get rate-limited.

The API Playground in your Dashboard now has Auto as a top-level Mode, alongside the Product row (Single, Browser, Proxy Finder). You can try the cost-aware ladder and session amortization end-to-end, set timeout_ms for the total budget, and pass ignoreProxies directly from the UI. Picking Auto switches off the per-product engines and vice versa.

Two controls from the per-product endpoints landed on Auto. timeout_ms (5s to 180s, default 120s) is now the total budget for the whole call; the orchestrator portions it across internal rungs so one sub-call can't eat your entire window. ignoreProxies lets you pass a list of proxy IDs Auto should skip, honoured across warm sessions, the good-exit probe, and the new-exit search.

Spanish, French, Japanese, Korean, Portuguese, and Russian joined the existing seven languages. The main site, the blog, the docs, and this Updates Portal all read in 13 languages now. Pick from the globe switcher in the header; your choice carries across surfaces via a shared cookie.

The Dashboard playground now has an Auto tab. Paste a URL, hit Send, watch Auto crack it. You can set your own success criteria via the Validate group, and a URL history typeahead suggests previously-used URLs as you type (one shared list across all endpoints, capped at the last 50). Non-applicable fields hide per engine, so each tab shows only the inputs it actually uses.

Every API response now carries an x-foura-credits header showing what that call cost. A new Budget view in the Dashboard breaks credit spend down by API key, domain, and client IP with a timeline chart, and the Playground shows the same on each response so you can size a workload before scaling it up.