Anthropic Acquires Stainless SDK Tool, Cutting Off OpenAI and Google

One in four professional software developers worldwide has used code that Stainless generated. Most of them didn't know. That's how API infrastructure is supposed to work: invisible, reliable, just there. As of May 18, 2026, Stainless is no longer there for most of the industry.

Anthropic announced that it acquired Stainless, the New York-based startup that turned API specifications into production-ready software development kits across TypeScript, Python, Go, Java, Kotlin, and more. The deal is reportedly worth more than $300 million, according to The Information, though Anthropic hasn't confirmed the price. What Anthropic did confirm is direct: the Stainless team is joining, the tools are being pulled in-house exclusively, and the hosted platform for all other customers is shutting down immediately.

The shutdown is already in effect. No new signups. No new projects. No new SDKs generated through the hosted Stainless platform. Existing customers retain full ownership of the SDKs they've already generated and can modify them freely. But the infrastructure that maintained those SDKs automatically, updated them when APIs changed, and kept documentation in sync? That's gone.

For developers at companies that relied on Stainless, the immediate impact is indirect. Your existing SDK still works. Your code still compiles. But your company now owns a maintenance problem that Stainless was solving quietly in the background, and that problem will become visible over time as APIs evolve and SDK libraries don't keep pace.

How Stainless Built the AI Industry's SDK Infrastructure

Stainless launched in 2022 with a specific, unglamorous goal: make SDK maintenance automatic. The pitch wasn't about AI models or fine-tuning or inference. It was about plumbing, and it addressed a problem that every API company eventually faces.

Developers don't just want a REST endpoint. They want a Python library that feels idiomatic to Python, a TypeScript package that handles types correctly, a Go module that follows Go conventions, a Java SDK that looks like it was written by someone who actually uses Java. Building those libraries is straightforward work. Maintaining them across six programming languages while your API evolves is relentless and tedious. One new endpoint means six library updates. One deprecated parameter means six deprecation notices across six codebases. One breaking change means six migration guides, six changelogs, six rounds of developer communication.

Stainless automated all of it. Companies fed the platform an OpenAPI specification, and Stainless generated polished, idiomatic SDKs across every major language, complete with documentation, type definitions, error handling, and automatic updates every time the underlying API spec changed. The generated code wasn't placeholder scaffolding; it was production-ready library code that companies shipped directly to developers.

Stainless also built MCP server tooling alongside its SDK generator. The Model Context Protocol, introduced by Anthropic as an open standard for connecting AI agents to external tools, requires connector infrastructure to work in practice. Stainless generated MCP servers that let AI agents use an API in a structured, tool-call-based way, not just raw HTTP calls. This is the part of Stainless's business that connects most directly to why Anthropic wanted the company.

The customer list Stainless built included companies that understood this SDK problem intimately: OpenAI, Google DeepMind, Perplexity, Groq, Replicate, Runway, Cloudflare, and Anthropic itself. Anthropic has used Stainless to power every official Claude API SDK since the earliest days of its developer platform. Alex Rattray, Stainless's founder, came from Stripe, where he helped build the payment infrastructure that made Stripe famous for developer experience. He brought that same focus to AI tooling when he started Stainless in 2022.

By acquisition day, roughly one-quarter of professional software developers globally had used a Stainless-generated SDK or documentation page without ever knowing it. That figure comes from Stainless's own announcement. It puts the acquisition in concrete terms: this wasn't a niche developer tool. It was shared infrastructure that underpinned the developer experience of multiple major AI platforms simultaneously.

The acquisition was announced on May 18, 2026. Reports of advanced talks had surfaced around May 14. The Information reported the price at more than $300 million; Anthropic declined to confirm or deny the number publicly. The Stainless team joins Anthropic's platform engineering organization and will focus on Claude-specific tooling. This deal is not structured as a product continuation under a new banner. The commercial product is being wound down. The team is being absorbed into Anthropic's internal platform engineering work.

Katelyn Lesse, Anthropic's Head of Platform Engineering, framed the acquisition around agent connectivity: "Agents are only as useful as what they can connect to." That sentence points directly at the Model Context Protocol. MCP is the open standard Anthropic introduced to give AI agents a standardized way to interact with external APIs, data sources, and tools. Stainless built the best tooling in the industry for generating those MCP servers automatically from API specifications. Bringing that capability in-house means Anthropic's platform engineering team now includes the people who have thought longest and hardest about the problem of connecting AI agents to external tools, and that expertise now benefits Claude's ecosystem exclusively.

Companies Affected and the Migration Path Forward

OpenAI is the most prominent company affected by the shutdown. It used Stainless to maintain official SDKs across Python, TypeScript, Go, Java, and more. Earlier reporting suggests OpenAI had previously tried maintaining those SDKs internally but found the cross-language maintenance burden high and adopted Stainless because automation made the process significantly more manageable. OpenAI has the engineering resources to rebuild this capability through in-house tooling, acquisition of a Stainless competitor, or contracting with one of the emerging alternatives. None of these paths are fast or cheap. The more significant problem is the transition period: during the months between the Stainless shutdown and whatever replaces it, OpenAI's SDKs will require manual maintenance, which means slower updates, more potential drift between API capabilities and client library features, and more documentation lag.

Google DeepMind is in a similar position with Gemini API tooling. Perplexity and Groq are smaller companies with more constrained engineering resources; for them, the shift from automated Stainless maintenance to manual SDK upkeep could have a more visible near-term impact on developer experience. Runway and Replicate are in similar positions, each relying on Stainless for developer-facing API libraries without the scale of larger labs. Cloudflare, a notable outlier as an infrastructure company rather than an AI lab, now needs to own a maintenance capability it previously outsourced.

Beyond the named companies, hundreds of smaller API businesses relied on Stainless without appearing in any announcement. The platform was designed to be invisible. Those teams face the same choices as the major labs, with less engineering bandwidth to solve them.

For teams working through this transition: previously generated SDKs are still functional and still owned by the companies that generated them. No running applications are breaking. The problems emerge over time as APIs evolve without corresponding SDK updates. Transition guidance is available at app.stainless.com/transition. Several alternatives have positioned themselves to capture displaced customers: Speakeasy supports TypeScript, Python, Go, Java, and Ruby SDK generation from OpenAPI specs and has been the most visible in reaching out to Stainless's former customers since the announcement. Fern focuses on SDK generation with strong TypeScript support and is expanding language coverage. liblab offers comparable core functionality across major languages. None of these alternatives yet demonstrates Stainless's breadth or depth of production quality, but the market vacuum Anthropic created is already accelerating their development.

The Stainless Acquisition and AI Developer Platform Competition

Anthropic acquiring Stainless fits a pattern building across the AI industry for the past 18 months: major AI labs are pulling shared infrastructure in-house rather than relying on third-party vendors for capabilities that touch the developer experience. A tool like Stainless looked, for years, like neutral ground. All the major AI labs paid for access to the same infrastructure, and Stainless served them without discrimination. That arrangement made sense when the AI industry was building the basic scaffolding of developer ecosystems. It becomes strategically untenable when developer experience is itself the competitive battleground.

Once Anthropic decided that SDK generation was strategic infrastructure rather than a commodity service, the choice was binary: own it exclusively or let a competitor own it. Anthropic chose to own it and remove it from competitors' reach simultaneously. The Information described this acquisition as part of a trend of AI companies pulling key suppliers closer. The precedent from other industries is instructive. Hyperscalers spent years building their own networking hardware, storage systems, and content delivery networks because dependence on shared vendors introduced competitive risk and operational cost. AI companies are running the same vertical integration playbook at a much faster pace, and developer tooling is one of the first fronts in that consolidation.

The competitive math is straightforward. Anthropic spent $300 million to acquire a tool that simultaneously improves its own developer platform and increases the maintenance burden for its largest competitors. At the scale of developer ecosystem competition, that trade looks efficient.

For developers choosing which AI platform to build on, the Stainless acquisition reinforces a pattern visible for a while: Anthropic treats developer experience as a first-class product decision, not an afterthought. The Claude ecosystem has made consistent, specific investments in developer tooling, including Claude Code for AI-assisted development, the MCP specification for agent connectivity, and now the in-house acquisition of the SDK generation team that underpinned the developer experience for the entire industry.

For a broader comparison of how Claude, GPT, and Gemini stack up on developer experience, model capabilities, and deployment options, our AI model comparison guide covers the key differences across the major platforms.

The competitive pressure this acquisition creates for OpenAI and Google is real and specific. Both companies have invested heavily in developer ecosystems. Both now have an SDK maintenance problem to solve during a period when developer attention is highly contested. OpenAI's Python SDK and TypeScript SDK are core to its developer relationships. Google's Gemini API SDKs are critical to its enterprise developer strategy. Neither company can afford for those libraries to visibly degrade while Anthropic's SDKs benefit from the team that automated this problem better than anyone.

What the acquisition doesn't change is what any AI model can do today. Claude's capabilities didn't improve because Anthropic bought Stainless. What changed is the infrastructure layer: who maintains the tools that connect developers to these models, and whether that maintenance is automated and high quality or manual and inconsistent. The developers who feel this most acutely aren't the ones using Claude. They're the developers at companies that have built products on OpenAI or Google APIs, relying on SDK quality that was quietly maintained by a startup that no longer exists in its original form. For those developers, the next question is whether their platform of choice moves fast enough to replace what Stainless provided.