Inkeep

#ICYMI: In a recent #RAG Brag episode, Nick Gomez, Founder and CEO of Inkeep shared how they use RAG and Pinecone to power search and chat experiences tailored for developers. Check out their story in more detail: https://hubs.ly/Q02H7Qm_0

Feeling the heat 📈

Introducing the new and improved Pinecone AI chatbot powered by Inkeep, an AI chatbot solution running on Pinecone. You can now find relevant search results and resources faster from across the Pinecone site — from our docs, blogs, or forums — and even open a support ticket. https://hubs.ly/Q02skpFF0. https://hubs.ly/Q02skpFF0

Love devex and curious about how LLMs and AI are changing the expectations of developers? Chatting with Nolan Di Mare Sullivan tomorrow about the intersection - including OpenAPI design, SDKs for GenerativeAI, and practical ways we've seen teams incorporate AI for their support, docs, and PLG workflows.

Leveraging your OpenAPI spec is crucial for building a great API. But OpenAPI isn’t easy to grok. That’s why we built an AI-powered OpenAPI reference, it’s now easy to answer all your questions 📘 The last couple of quarters, our team at Speakeasy has developed extensive OpenAPI knowledge, which we’ve compiled into a 📚 "Complete OpenAPI Reference" 📚 that demystifies the specification for developers. What sets our OpenAPI Reference apart? Here are a few highlights: 🤖 AI-powered search and chat to make answering your questions easy. Built using our friends Inkeep and Codehike 💡 Example-rich so that you can learn from how others write their spec 📘 Comprehensive coverage of the OpenAPI 3.1 Specification and where it differs from 3.0 🌍 Open Source - this documentation is open source and available for experts from everywhere to contribute to! Our reference guide is designed to be your go-to resource, whether you're just getting started or are looking to deepen your existing knowledge. Dive into the details and join us in shaping the future of API development. #OpenAPI #APIDevelopment #OpenSource #Collaboration Check out the site in the comments below 👇

In March we answered 120,000 end-user questions with AI. Here's what we've learned getting to this scale. <<not an April fools post>> 💬💬 Users like the conversational nature. Over 40% of the questions we get are follow-ups. 🏃 Users ask questions at every stage of their lifecycle. 1. "Is it possible..." - from users first evaluating the product (prospects) 2. "How to..." - from users looking for scenario-specific guidance 3. "Help me fix this..." - from users looking for troubleshooting help 🤝 "Meet users where they are" holds true. We've seen companies start with us just on their Discord or Slack, then on their docs, and then in-product as part of their support center. For these cases, total usage has grown with each new integration. 💡 Developers in particular prefer to self-troubleshoot if they can. We see teams get 10-20x number of questions from AI chat compared to e.g. Discord or support tickets. While some developers prefer to read the docs E2E or reaching out to maintainers, many have come to expect immediate, tailored guidance as a first pass. -- I think these patterns show why "AI chat" is fundamentally different than just a new form of search. Users depend on it for more than navigation, ask richer questions, and interact with it in all types of channels, not just a search bar.

📢 We're hiring! Looking to work in a fast-paced, fast-growing startup building AI-native full-stack experiences? Position: Founding Engineer Background: Full-Stack // Backend JavaScript development Location: NYC 🗽 (in-person) or SF About us: - We help companies turn their content into AI search and support copilots - 100+ teams use Inkeep to answer 100k+ questions/mo with AI - Seed stage, investors include Y Combinator and Khosla Ventures - Team of 5; founding team are MIT alum What we're looking for: - 3+ years of experience - Strong API, SDK, TypeScript, and database design skills - Experience with Next.js or other full-stack JS frameworks - Excellent engineering, product, and devex sense - Likes shipping and iterating quickly Really nice to haves: - Experience with DevOps and CI/CD - Fluent with web fundamentals (Web APIs, React, Server <> Client interactions, JS runtimes, etc.) - Familiar with mono-repo, package management, and build tools - Open-source contributions

What does app development look like in a world powered by Generative AI? In-product, support, and chat-based copilots are great (it's what we do at Inkeep), but there are many non-obvious ways in which GenAI will change what users come to expect out of modern applications. Recently, I used Vercel's AI SDK to create dynamic, real-time LLM-powered reporting for Inkeep. A key scenario for us is helping teams identify the top documentation and feature gaps in their product by analyzing AI chat conversations. Teams actively use these to get actionable visibility on how they can improve their documentation and product. 🕒 Before: Our first shot at this was weekly and monthly reports. While these were handy, it became apparent that teams want to view these insights when they're most relevant, like when they are doing roadmap or sprint planning. ⚡Now: Teams can generate a report at any time, with the categories they care about, for any time period. These on-demand reports leverage the Vercel SDK to stream the report -- right as it's being generated. No waiting around. I think this is the future of app development: hyper-contextual, intelligent, dynamic experiences + UIs that can delightfully represent complex state transitions. Chat-based UIs are just one part. All to say, I think it's actually the most exciting time to be a designer, PM, or app developer. The "easy", repetitive stuff will be highly automatable - leaving room for more creative thinking about what's possible with UIs. s/o Guillermo Rauch and team for providing a foundation for this transition.

AI chatbot on Oso docs lookin fly. Ask it anything about Oso or authorization generally h/t to Nick G. and the Inkeep team for making this happen in record time

It's been a pleasure working with Nick G. and the Inkeep 🤝 team. They've been great partners helping us get live with powerful AI search and augmenting our docs experience in just a day. See for yourself, link in comment 👇