Latest updates and improvements to FourA
Send faster than your plan's per-minute limit and requests up to that limit still go through, since refused ones don't count toward the minute. Retry-After gives the seconds until the next one passes, and refusals arriving at over twice the limit get a 30-second pause. New proxy port connections are counted the same way.
FourA MCP 0.7.3 publishes tool schemas that validate under both JSON Schema draft-07 and 2020-12, so clients with a strict 2020-12 validator list and call all four tools. What a call accepts hasn't changed. Update @fouradata/mcp, or connect to the hosted server, which already runs 0.7.3.
A request refused because it hits a limit on your own plan doesn't add to your traffic total. Responses from the target site still count, since that traffic was carried.
Traffic figures on your Dashboard now read in gigabytes wherever a gigabyte limit sits beside them, so a number and its ceiling always compare directly. Anything under a gigabyte still shows in MB, and large figures drop decimals that add nothing.
Every top-up on your Billing page now has its own line: what you bought, what's been used from it and what's left. Usage draws from the top-up that expires first, and the usage card lists your included traffic and your top-ups separately.
When an Auto call hits a limit on your plan, Activity now shows it as Rate Limited, and the response carries the same X-FourA-Limit header a direct call returns. You can tell which limit you reached at a glance.
When a target's hostname doesn't resolve, Single, Browser, Proxy Finder and Auto now answer the way they do for any site they can't reach: status 0 and an error naming the host. It's recorded as a target outcome (application_fail), while private-network targets still get the 400 refusal. Outcomes reference →
Open Activity right after an Auto call and the row says its attempts are still settling, then fills in the attempt list, credits and bytes in place a few seconds later. No reload, so the row you're watching stays where it is.
Your Dashboard now counts each Auto call as one request, and the Single, Proxy and Browser attempts behind it are listed separately by name. Overview, Usage & Limits and Activity all carry the Auto product, and success and latency are measured on your calls.
The benchmarks page now tells you when the published run is more than 45 days old, read against your own clock, and points at the next monthly run. Sites change their defences without notice, so an old run is history, not a current reading.
The plans page now shows how much proxy port traffic each plan includes, from 2 GB on Free to 750 GB on Business, and what an extra gigabyte costs. Both figures also sit in llms.txt for agents that read plans from there.
A consumption limit you've reached, such as credits, traffic, Browser pages per day or rate, now shows a red At limit, because the next request is refused. A count limit at its ceiling, such as API keys or team members, shows amber, since it only stops the next addition.
The Billing page now shows the same four cards on every plan: credits or standard traffic first, whichever the plan is sold by, then the other figure, Premium, and the Proxy port. Every card has a Buy button, and where your plan doesn't price that item it says so and offers a contact link.
Your Limits page gets a Traffic panel with the total, the API, the proxy port and premium, each with a bar wherever your plan sets a ceiling. The API keys table adds Traffic and Premium columns, and every product panel on the Overview shows its own traffic.
On phones and tablets, the menu on the main site now opens and closes the moment your finger lifts, iPhone included. It fills the screen in two columns: Products, Pricing and Tools on the left, Resources and Compare on the right, under headings that match the footer's.
