Microsoft confirmed a deprecation date this week that's easy to skim past: the SOAP and POX/HTTP APIs for Bing Webmaster Tools stop working on August 31, 2026, per Search Engine Roundtable's recap and Microsoft's own developer documentation. If you're not sure whether that sentence applies to you, it's worth two minutes to find out, because the answer, for a surprising number of teams, is yes.
What's actually retiring in Bing Webmaster Tools
This kind of notice tends to get filed under 'developer housekeeping' and skipped by anyone who isn't directly writing the integration code, which is understandable and also exactly the failure mode worth avoiding here. Deprecation notices for legacy protocols rarely make headline SEO news, precisely because the migration itself is small and boring by design. That's also why they're easy to miss until the deadline has already passed and something quietly stops updating.
Bing Webmaster Tools has offered two ways to pull data programmatically for years: the newer JSON/HTTP, or REST, API, and the older SOAP and POX/HTTP protocols that predate it. Come August 31, 2026, the SOAP and POX/HTTP endpoints stop working entirely. Anything still calling them, a legacy internal script, an old reporting dashboard built years ago and never touched since, a piece of an agency's client-reporting pipeline inherited from a previous hire, goes dark on that date with no grace period mentioned in the announcement.
The replacement isn't a redesign. Per Microsoft's own migration notice, every API method that existed over SOAP/POX remains available over JSON/HTTP with identical functionality. Existing API keys carry over without needing to be reissued. Quotas, rate limits, and permissions are unchanged. This is Microsoft retiring a legacy transport protocol, not changing what data is available or how access is governed.
Why this matters more than it sounds like it should
Bing gets a fraction of the attention Google Search Console does in most GEO and SEO reporting stacks, and that's mostly rational; Bing's organic search share is far smaller than Google's for most businesses. But Bing's index has taken on outsized importance for a reason that has nothing to do with Bing's own search market share: it's a meaningful part of how ChatGPT retrieves live web content. We wrote about a concrete case of this just above in today's batch, where a site with essentially zero Google visibility built real ChatGPT citation volume entirely through Bing's index, until Bing deindexed it on August 3.
It's worth being precise about the scale of that reliance rather than overstating it. Bing is not the majority of ChatGPT's retrieval pipeline, and plenty of ChatGPT's answers draw on trained knowledge or other sources entirely, especially for topics with less time-sensitive demand. But for live, current-web retrieval specifically, the architectural relationship between Microsoft's search infrastructure and OpenAI's models is well established and long-running, dating back to Microsoft's early investment in OpenAI. That means Bing's index health is not purely a Bing-market-share question. It's a GEO-infrastructure question, whether or not your team has ever run a Bing ad or thought about Bing search traffic as worth optimizing for directly.
That's the context this API retirement lands in. If your team runs any automated monitoring of Bing indexing status, crawl activity, or search performance, whether for classic Bing search reporting or as an early-warning system for the kind of Bing-index dependency that case study just illustrated, and that monitoring was built against the SOAP or POX/HTTP endpoints, it stops working on August 31 with no visible error beyond a data feed going quiet. A quiet failure is worse than a loud one, because nobody investigates a dashboard that simply stops updating as urgently as they investigate one throwing errors.
| SOAP / POX/HTTP (RETIRING) | JSON/HTTP REST (CURRENT) | |
|---|---|---|
| Status after Aug 31, 2026 | Stops working | Fully supported |
| API method coverage | Full | Identical — full parity per Microsoft |
| API keys | Existing keys used | Same keys, no reissue needed |
| Quotas / rate limits / permissions | As previously set | Unchanged |
BingBot has its own exposure to today's Cloudflare news
There's a second reason to have working, current visibility into Bing's crawler behavior right now, and it's a direct connection to Cloudflare's new AI crawler defaults taking effect the same month, September 2026. Cloudflare's policy names three crawlers explicitly as 'multi-purpose,' serving more than one function for their parent company: Googlebot, Applebot, and BingBot. All three get the same treatment under Cloudflare's most-restrictive-rule logic. A new domain onboarding to Cloudflare after September 15 that blocks AI training crawlers, without explicitly carving out an exception, doesn't just risk losing Googlebot. It risks losing BingBot too, for the identical structural reason: Microsoft runs BingBot for both classic search indexing and AI-related purposes tied to Copilot, and Cloudflare's policy can't tell the two apart any more than it can for Google.
That makes this a month where Bing's infrastructure deserves more attention than its search-market share alone would justify, for two unrelated reasons landing almost simultaneously: an API you might be quietly depending on stops working August 31, and a crawler-blocking default that can silently take BingBot down with GPTBot takes effect September 15 for anyone launching a new property on Cloudflare. Neither is a crisis. Both are the kind of quiet, compounding infrastructure change that's cheap to handle proactively and expensive to diagnose after the fact, once a report has already gone stale or a citation number has already dropped.
The Bing Webmaster Tools migration itself is small
Because the underlying functionality and access model are unchanged, the actual migration is a code-level swap: point existing calls at the JSON/HTTP endpoints instead of the SOAP/POX ones, using the same API key, and update whatever response-parsing logic assumed an XML/SOAP payload to handle JSON instead. Microsoft provides a migration guide alongside the deprecation notice for exactly this. For a team that built its own reporting scripts, this is an afternoon of engineering time, not a project.
The bigger pattern this fits
This kind of deprecation is exactly the sort of thing our own guide to reading Google Search Console like an analyst argues teams should build institutional habits around: know precisely which first-party data sources your reporting depends on, keep that dependency list current, and treat a vendor's deprecation notice as an action item with a deadline, not background noise to skim past. Google's own Search Console has gone through comparable platform changes this year, including the Platform Properties rollout that changed how creators track cross-platform content performance. Infrastructure that reporting pipelines depend on keeps shifting under them, on each vendor's own schedule, and the cost of missing one of these notices is never the deprecation itself. It's the multi-week gap between a report going stale and someone noticing.
None of this is a reason to overhaul anything beyond the specific endpoint swap. It's a reason to actually run the audit this week rather than filing this piece under 'not urgent.' The Bing-deindex case study we published today is a good reminder of the stakes on the other side of this exact infrastructure: teams that don't have first-party visibility into Bing's index status find out about a problem only when a downstream AI-citation number moves, weeks after the actual cause. A five-minute API migration, done now, keeps that visibility intact.
If your team has never once logged into Bing Webmaster Tools, this is also a reasonable week to start. It's free, it takes a few minutes to verify a domain, and it's the only first-party window most sites have into whether Microsoft's crawler infrastructure, the same infrastructure feeding a meaningful share of ChatGPT's live retrieval, can actually see the site at all. Given everything above, that's a cheap insurance policy against finding out the hard way, the way this week's other case study did.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Josh leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.