Lorenz Wieseke’s Post

Most automation builders waste dozens of dev hours wiring separate API routes for every single action their AI needs to take. Create event? One route. Update sheet? Another route. Search calendar? Yet another. It's tedious, expensive, and painfully slow. I built an 𝗠𝗖𝗣-𝗽𝗼𝘄𝗲𝗿𝗲𝗱 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄 that lets Claude (or any LLM) call n8n tools directly through chat. No manual route building. No separate integrations. Here's the system: → Chat command sent to AI Agent → MCP Client routes to Claude → Claude picks the right tool automatically → MCP Server Trigger fires the action → Google Calendar/Sheets updated instantly The coolest part? The LLM auto-generates parameter schemas and figures out which tool to use from plain conversation. Type "add Roy with $500 budget to the sheet" and it just works. Result: One chat interface controls unlimited tools without building separate routes for each action. This is early tech, but once MCP fully matures, it'll cut integration build time by 80-90%. Want to try it? 1️⃣ Comment "𝗠𝗖𝗣" below for free access to the workflow 2️⃣ Make sure we're connected so I can send it over #n8n #WorkflowAutomation #AIAgents #NoCode #AutomationBuilders

AI models are great at reasoning, but they don’t actually know anything about your systems. The Model Context Protocol (MCP) changes that. It gives models like GPT and Claude a standard way to talk to tools, APIs, and data, without custom integrations or vendor lock-in. If you think of USB-C as the connector that unified hardware, MCP is doing the same for software. Here’s how it works and why it matters for anyone building connected, context-aware systems: https://lnkd.in/dQrQhnrv

𝐑𝐞𝐩𝐨𝐬𝐢𝐭𝐨𝐫𝐲 𝐑𝐚𝐝𝐚𝐫 𝐏𝐑#𝟐𝟎 - 𝐎𝐩𝐞𝐧 𝐒𝐨𝐮𝐫𝐜𝐞 𝐀𝐈 𝐈𝐧𝐟𝐫𝐚 𝐈𝐧 𝐌𝐨𝐭𝐢𝐨𝐧 Back with your no-fluff scan of open source software infra. This week Claudius and myself highlight how OSS agent stacks and AI infra frameworks are scaling. LangChain’s 125m USD round shows that open frameworks can evolve into real platforms, while Immich, Pathway, and Polar keep expanding what’s possible for self-hosted data, live pipelines, and developer monetization. 𝐖𝐡𝐚𝐭’𝐬 𝐫𝐢𝐬𝐢𝐧𝐠 𝐟𝐚𝐬𝐭: ⚡ LangChain – Raised 125m USD at a 1.25b USD valuation led by IVP, proving that OSS-first frameworks can scale even in thin AI infra layers. 🖼️ Immich – Self-hosted photo and video management with AI search, tagging, and multi-device sync. 🔎 Pathway LLM App – Template-based RAG and live-data AI pipelines for enterprise search and indexing. 🧠 Wave Terminal – Cross-platform terminal combining CLI, editor, browser, and AI assistant in one workspace. ⚙️ Agent Lightning (Microsoft) – Reinforcement learning trainer for AI agents supporting multiple frameworks. 📚 Open Notebook – Multi-model, self-hosted research notebook for PDFs, audio, and contextual chat. 💳 Polar – Open-source payments and monetization platform for developers selling SaaS and digital products. From OSS frameworks to agent training and monetization infra, this wave shows that open AI infra is not only research-grade but becoming commercial and production-ready. 📡 𝐏𝐑#𝟐𝟎 𝐟𝐮𝐥𝐥 𝐰𝐫𝐢𝐭𝐞𝐮𝐩 𝐡𝐞𝐫𝐞: https://lnkd.in/gUDGSZRc 👩💻 𝐎𝐒𝐒 𝐇𝐞𝐫𝐨𝐞𝐬: Harrison Chase (LangChain), Zuzanna Stamirowska (Pathway), Michael Sawka (WaveTerm), Xufang Luo (Agent Lightning), Birk Jernström (Polar) #opensource #AI #LLM #RAG #agentic

𝐎𝐩𝐬 𝐰𝐢𝐭𝐡𝐨𝐮𝐭 𝐃𝐚𝐬𝐡𝐛𝐨𝐚𝐫𝐝𝐬 Leads, tickets, and payments don’t need “a place to check”—they need to come find you. — Webhook ingests → enrich with Clearbit-like data → route by ICP score. — If value ≥ threshold, auto-create a 3-line brief + DM the owner. — If value < threshold, archive with a human-readable log. Outcome: fewer tabs, faster replies, happier P&L. Ops isn’t another tool; it’s decisions pre-made. #n8n #Automation #NoCode #WorkflowAutomation #APIs #Ops

1979: “Computers can’t be accountable.” 2025: Agents are not there, yet either. Let them get the context. You make the call. The Nimble Lab is using the Gemini 2.5 Computer Use model to power out agents. The new model allows interaction with user interfaces. It's powerful. We deploy 𝗪𝗲𝗯 𝗦𝗲𝗮𝗿𝗰𝗵 𝗔𝗴𝗲𝗻𝘁𝘀 in high scale to navigate, find, and validate the dat across the web. Our platform converts the logic into a trusted workflow. Agents can not be accountable to do this work in zero shot: → read live pages → validate across sources → simulate outcomes → request approval for high-impact moves 𝗪𝗵𝗮𝘁 𝗸𝗲𝗲𝗽𝘀 𝘀𝗽𝗲𝗲𝗱 𝗮𝗻𝗱 𝘀𝗮𝗳𝗲𝘁𝘆: → Role and intent checks: “Who is asking? For what?” (no shadow requests) → Multi-source consensus: 2 to 3 independent signals before action → Set proper evels and guardrails. → Reversibility and logs: every agent action is explainable and reversible Let agents 𝗽𝗿𝗼𝗽𝗼𝘀𝗲 price/promo/shelf updates; let humans 𝗱𝗲𝗰𝗶𝗱𝗲 (approve, tweak, or block). You get the upside of automation without handing over the P&L. As a side note, I think OAI new agent builder is a product with a lot of potential, but agents still need quite a lot of debugging before taking full responsibility. Which action in your stack is safe to auto-approve today, and which should always require a human check? 📌 Let's connect on Slack: https://t2m.io/WbxCy8S

Hello, happy Monday! We often imagine “adding AI” means reinventing our entire stack. But what if you could bring intelligence into your backend, without rebuilding it? This week in my weekly tech series on Medium about how I engineer impacts, I explored a pragmatic approach for modern engineers: - Why AI Middleware is the missing layer between APIs and systems. - How to design a “reasoning layer” that learns context and makes smarter decisions. If you’re building enterprise APIs, integrations, or just experimenting with AI; this one's for you. Have a quick read, here: https://lnkd.in/gJ-eUjuh Always here to hear you if you have worked on similar challenges or for a simple chat over a cup of coffee ;) Hashtags: #AIEngineering #BackendDevelopment #SystemDesign #SpringBoot #EnterpriseAI #Middleware #SoftwareArchitecture #FullStackDevelopment #APIs

Building Your First MCP Server   The best way to understand MCP? Build something with it. Key Steps: - Initialize the Server: Define your server name, version, and capabilities - Expose Resources: List available files/directories as resources - Implement Tools: Create functions for reading, searching files - Handle Requests: Process incoming requests from MCP clients - Connect to Transport: Use stdio for local or SSE for remote connections - Testing Your Server: Connect it to Claude Desktop and watch it work! Resources: Official SDK: https://lnkd.in/gWDGY_kz Examples: https://lnkd.in/gusuDYy4 Docs: modelcontextprotocol.io The barrier to entry is incredibly low. If you can write a REST API, you can build an MCP server. What would you build first? 🤔 #Coding #MCP #AIIntegration #DeveloperCommunity

Breaking down complex digital tools doesn't have to feel like decoding ancient scripts. The secret: simplify. Turn big features into small, clear steps. When users know the "how" and "why," they trust. At Tech TLC, we make training that clicks.

Building RAG Agents in Minutes — No Code, No Hassle! In my latest workflow, I explored a faster, smarter way to build #RAG (Retrieval-Augmented Generation) agents using #n8n + #Pinecone Assistant — and it completely changes the game! What’s new? No preprocessing pipeline. No manual chunking. No custom embedding flow. Just drop your documents in, and Pinecone takes care of the ingestion, splitting, and indexing on the backend — automatically. I connected it all together to get: Accurate, context-aware answers. Page-level citations. Exact text quotes straight from the knowledge base. Load your docs → Connect in n8n → Test grounded responses → Deliver transparent, trustworthy answers. #n8n #Pinecone #AIagents #RAG #Automation #NoCode #AIAutomation #OpenAI #Gemini #MachineLearning #AIWorkflows

Super excited to share this! I just built a small but mighty AI Agent using n8n that takes your answers from a Tally form, analyzes them instantly, and sends you a smart summary right to your email inbox! ⚡ This was such a fun project to build — simple idea, but the automation possibilities are endless. Try it out here 👇 🔗 https://tally.so/r/mKqDM8 Big thanks to Youssef Elbadry and Instant Software Solutions Software Solution for the amazing support and motivation! 🙌