Pipedream

I am excited to announce that Pipedream has entered into a definitive agreement to be acquired by Workday. Our mission to enable customers to build powerful, connected, and at-scale AI agents will accelerate as we bring together: 1. Our AI agent-building products and 3,000+ connectors 2. Workday’s leading enterprise customer base and 75+ million users under contract. As we launched String.com and Pipedream Connect this year, it became clear AI agents will: - Touch every team - Impact every product - Integrate with every software within the enterprise We are excited to build on these insights and successes to have an impact at a greater scale. An enterprise scale. A global scale. This journey began with a humble MVP, followed by years of iteration based on feedback, engagement, and contributions from our community. Thank you to our customers, partners, friends, and investors for your support and trust. Thank you for pushing us to build faster and be more ambitious. We're excited to soon begin this next chapter with Workday after the transaction closes. I’ve studied Workday deeply over the last few months, and it’s an awe-inspiring, at-scale, and value-driven business. I am thrilled to work with Aneel Bhusri, Carl Eschenbach, Gerrit Kazmaier, Gabe Monroy, and Peter Bailis to expand the leading AI platform for managing people, money, and agents. See the press release in the comments, and more details soon. Now, it is day one again. Back to building!

Here's something that sounds impossible in SaaS - We have 1 person handling support for over 30,000 active users. In reality, I think we're less than 2% into our automation journey. For context, I see 3 macro trends happening right now: 1. Automating the parts of your business that you pay humans to do. This has always existed, but with AI agents, it's 10x more relevant. 2. Companies doing far more revenue per employee. When I ran my first startup, success was unfortunately measured by headcount (and money raised). Now it's revenue per employee - literally the opposite. 3. What we call 'vibe automating' - using natural language to build AI agent workflows. At Pipedream, we try to live these trends. Since we sell an automation product, there's almost no part of our business that doesn't have some element of automation. Every communication you get after signing up is sent by our product. Every support interaction is automation first. Our lead scoring - all automated. (Our newsletter is one of the only things that’s not on autopilot.) Just this week, we built an agent to handle complex questions that our automated system can't. Instead of customers talking to a bot, it comes from human - but the research and response prep is automated. We basically ask: "How can we make our job easier?" That's the first question to ask in the AI agent automation journey. We keep finding areas we can automate - which in turn helps us improve our product. This is what 2% automation looks like.  The other 98% is where it gets interesting.

It's been 2.5 months since we launched String to the public. We’re now growing faster than any time in our history. But it all started when I asked my team to answer 2 questions: 1. Can we build agents with natural language? 2. Can we dynamically code tools that we don't have pre-built? I put one engineer on each question. My theory was pretty simple: If we could solve both problems, our product would be 10x easier to use than Zapier AND solve 10x more use cases. 10x easier × 10x more use cases = 100x larger TAM (total addressable market). The 6-week research spikes proved it was possible. By February, we knew we could do both. So we put half the company on it. - 3 months heads-down - First internal version went live in May - Public launch in July Since then, it’s been up and to the right. Here's the funny thing though - we knew when we launched that this wasn't the product we ultimately wanted to build. String in its current form is an Alpha product. It says so right on our website. What we’re really building towards is fully non-deterministic AI agents. In the future, the entire agent IS the LLM. It's involved at every step, it adapts and makes decisions automatically. This is where workflows will become truly agentic. More expensive to run? Sure. But we’re betting on LLM costs going down and tool capabilities going up. We’re not limiting the product because it's expensive today. String taught us the lessons we needed to learn. The next version will make no-code builders a relic of the past.

Big news for developers building on Pipedream Connect! We just shipped two major upgrades that make it even easier to integrate Connect into your apps and AI agents: 1. New SDKs for Python, TypeScript, and Java With new SDKs for Python, TypeScript, and Java, you can get started in minutes instead of days. No need to reinvent auth flows or wrangle API requests (just install the SDK and focus on your app.) 2. All-new interactive, AI-friendly documentation Our completely redesigned documentation is interactive, AI-friendly, and built for how developers actually work today. Search, copy, or feed docs directly into your coding agent or IDE to unblock yourself instantly. We’re laser-focused on making Connect the easiest way to integrate 2,800+ APIs into your applications. Dive into the new docs here: pipedream.com/docs

Looking to take your String.com prompts to the next level? Check out this article by our Head of Revenue, Danny Archer on How to Write Great Prompts for String. He shares expert advice on building powerful, cost-effective, and reliable AI agents using String (Pipedream’s natural-language AI automation platform.) 3 takeaways you’ll learn in the article: - Why clear triggers are the foundation of every strong prompt - How to structure actions to avoid errors and cut costs - The importance of naming apps explicitly for faster, more reliable automation Try String today for free! Check out the full article here: https://lnkd.in/eKy2pTPY

🔥 Last week Pipedream just dropped Pipedream MCP, along with a demo app showing how easily you can plug their remote server into your AI app — letting you connect to everything in your stack through 2,700+ APIs and MCPs in seconds. Seriously, go try it now. Oh, and it’s open source — ready for you to fork, tweak, and turn into your own AI app today. What a world! 🤔 So, what are you going to build? 🔗 https://lnkd.in/ej_MqiYK 🔗 https://lnkd.in/eHGN_25D #pipedream #mcp #ai #producthunt

Thank you to Adam Silverman (Hiring) for sharing our Pipedream MCP Chat App. MCP Chat lets you talk directly with all the software tools you use and is now open source! Powered by Pipedream Connect, MCP Chat provides access to 2,700+ APIs with 10k+ tools via MCP. Check it out for yourself today --> chat.pipedream.com

Product Hunt Launch Day! We are excited to release MCP Chat today on Product Hunt. MCP Chat enables you to chat with your software across 2,700+ APIs & MCP servers with 10k+ tools. It includes built-in authentication management and tool discovery, and works with any LLM or framework. All the integrations are built by our team and battle-tested, with secure user auth built-in. On top of that all, it is open source (MIT license) so you can use it as a reference implementation to integrate Pipedream Connect into your app. Try Pipedream Chat for free today! Chat.pipedream.com

“We’ve spent over $1M building about 250 integrations over the past 2 years. And now, with Pipedream, we’re immediately offering 2,500." - Flo Crivello, CEO of Lindy Lindy was building integrations one-by-one to power their AI agents. They started using Pipedream Connect which instantly unlocked 2,800+ integrations without writing custom code. Now, Lindy’s agents can automatically take action across tools like Salesforce, Slack, Notion, and Gmail. Lindy’s agents just got a whole lot more powerful. Check out how they did it: https://lnkd.in/e6xgJKug

With over 2,700 integrated applications, the use cases for String.com are limitless. Thank you to Taj S. for sharing what you've built!