Stainless

Congrats to Dedalus Labs on the $11M seed round 🚀 Their open-source agent SDK (https://lnkd.in/eubtY-si) uniquely supports model handoffs, streaming, and tool calling in just one copy-paste. Excited to support the the team as they build the next layer of agent infrastructure.

🚀 New: Remote MCP servers Add --transport=http to deploy your MCP server over Streamable HTTP. Perfect for Claude.ai, LangChain, or any web-based agentic flow. https://lnkd.in/eFKv_zFp

We're excited to welcome Alex Lake as Account Executive ✱ Alex joins us from Pare AI, a company he founded that aimed to give every account executive their own personalized AI deal strategist. Prior to Pare, he worked at Common Room where he helped lead the company's enterprise GTM motion.

turbopuffer went from 2 hand-written SDKs → 5 Stainless-generated SDKs. ✱ TypeScript & Python clients 5-50% faster in real-life workloads ✱ Async Python SDK processing 100MB+ payloads ✱ Node query_scale: 3,436 ms → 2,778 ms (1.24x faster, no event-loop stalls) https://lnkd.in/dA-MkQi8

APIs are the invisible infrastructure that powers every email you send, every payment you process, and every login you complete. They're the handshakes between programs that make the internet work. Now AI needs its own kind of plumbing. Alex Rattray, founder of Stainless, joined Dan Shipper on the latest episode of AI & I to talk about MCPs (model context protocols)—the servers that let language models plug into software and get work done. The challenge: They don't work as well as they should yet. At Stainless, which builds robust SDKs that turn APIs into developer-friendly tools, Rattray has built MCP connections to Notion and HubSpot. He can ask: "Which interesting customers signed up last week?" The system queries the database, cross-references HubSpot, pulls notes from Notion, and delivers a summary. It's not flawless yet. But it's transforming work that normally requires multiple logins and manual searches. If you're building with AI or just trying to understand how it will integrate into your workflow, this conversation is worth your time.

MCP servers generated by Stainless now ship with a live docs search tool 🔍 If you were using an MCP server in Claude Desktop or Cursor and asked, "How do I call this model with a custom payload?" The AI agent could only rely on its pre-training or try to infer usage from the OpenAPI schema. That often leads to errors or hallucinations, especially with more complex APIs. The new docs search tool gives the MCP server a live, searchable documentation layer built directly from your project's latest OpenAPI spec and Stainless config. Here's what happens: 1️⃣ LLM queries the docs tool 2️⃣ MCP server responds with Markdown snippets - concise explanations, usage notes, and code examples pulled from the Stainless-generated SDK 3️⃣ LLM uses the retrieved info to answer or act https://lnkd.in/e-mikUbz

Dan Shipper shares how Every Inc. is going deep on multi-agent workflows across their products: 📦 Cora spins up subagents to debug UI in real time with Puppeteer 🧠 Spiral uses design + review agents to autonomously iterate until the UI matches the Figma mockup Panel discussion: https://lnkd.in/e8xqS2F4

ICYMI: Support for the new Conversations API is now available in the OpenAI SDKs, generated with Stainless. Conversations provide a flexible primitive for managing model context over multiple API requests. After a conversation is created, you can add or remove items in it. Conversation items can be any valid Responses API item type (like assistant messages, input images, tool calls, and more). No need to juggle a large input array over multiple turns in your own code. Learn more about the Conversations API in the OpenAI docs: https://lnkd.in/ee7NKCrS Check out OpenAI's official TypeScript SDK here: https://lnkd.in/epZB4GWY

Here, an image generation API exposes a single endpoint that accepts different model names and parameters. But as the API evolves, each model will require its own config. E.g. some ratios only work with certain models, and some models drop the seed parameter altogether. This requires a design choice in how you expose model-specific parameters. There are four concrete API shares to consider, plus a "default model" variant: 1️⃣ Common-denominator schema 2️⃣ Discriminated unions 3️⃣ Separate endpoints at /v2/generate_image/{model} 4️⃣ Separate endpoints at /v2/models/{model}/generate_image 5️⃣ Add a default model endpoint We dive into the trade-offs of each 👇 https://lnkd.in/eQC-4-gT

We're excited to welcome Kevin Whinnery as Head of Marketing & DevRel ✱ Kevin joins us from OpenAI, where he worked closely with Stainless on SDKs and docs. With 15+ years across marketing, DevRel & eng leadership, he'll lead our efforts to support the developer community building on Stainless.