For my #gamer friends out there, I'm making a game. If you like snarky #synthwave #2D #pixelart spoofy sci-fi #roguelites, well, this is for you. If not, check it out anyway. Link to my @Patreon in the reply. Help me shape what this will become.

MCP was adopted amazingly fast by the AI community. Everyone jumped on this bandwagon and for good reason. It's the connective tissue between agents, our app, and the external world. But MCP token usage can get out of hand. The silver lining - there are open source tools like MCPJam that are tailor made for those of us working on MCP services to clean up the bloat and serve better, faster, and richer experiences. Projects like OpenRAG, Langflow, and anyone deploying MCP services can all benefit. Come, listen to my conversation with Prathmesh Patel, CEO of MCPJam, and Mike Fortman, lead developer for OpenRAG, as Prathmesh shows us how MCPJam can help and we discuss some new trends in the MCP space. Pick your poison, then listen to the antidote: Links in reply. #MCP #AI #opensource #developer #openrag #langflow #mcpjam

Even in this age of coding AI's, SDK's are still your friend when building apps. Sure, you can point an agent to some online docs and say "build this into my app", but giving it clear instructions, code examples, and authoritative references is a much better way to ensure your agent does the right thing. Did you know that the OpenRAG GitHub repo contains skills for both the SDK and install? Point your agent to: Link in reply Have it save both openrag_sdk and openrag_install skills. Then hook OpenRAG into your app. #ai #sdk #mcp #openrag #applications #development #install #code

I've been working a lot with AI and RAG platforms for a while now, building lots of things. Most are the stock kinds of applications you might expect from an AI/RAG solution, but I've been experimenting with something different. I'm an avid gamer and I really enjoy finding ways to bring gaming, RAG, and AI together. I wrote a blog post about it and I would love your thoughts. Would you play this? #openrag #gaming #ai #rag #application

So, I find myself looking for more and more ways to reduce token count, especially when dealing with some of the more expensive providers. Nothing like blowing past an hourly or weekly limit in the middle of an #AI #coding session. 🤯 And lets be honest, #MCP is not helping. MCP #bloat is a thing. There are lots of MCP services out there that take a #YOLO approach to token usage. Don't be one of them! 😬 On our next episode of "The Flow" AI podcast, we'll put @Prathmeshp_16 CEO of @mcpjams, to the test as he shows us how to use MCPJam to keep our services from getting out of control. Join us live on May27th @ 1PM EDT! Link in reply.

Ever spent 20 minutes scrubbing through a video trying to find that one thing the speaker said? I have. Many, many times. Video is incredible for learning, but completely broken for reference. 🫠 So I built something to make my life easier. → Auto downloads YouTube videos (single, playlist, or entire channels) → Transcribes locally with Whisper Docling → Feeds everything into OpenRAG for semantic search The result? Ask a question in plain English, get the exact timestamp with full context back in seconds. I threw 12 hours of podcast episodes at it 🎙️ Searching through that content now takes seconds instead of hours. Everything runs locally too — no API keys for transcription, no shipping your data to a third party. @ollama OpenRAG handles the rest. If you sit on a pile of video content and wish you could actually search it, the blog post has the full breakdown plus a video walkthrough. Link in reply. OpenRAG is still in open preview at v0.4.1 — feedback and contributions welcome! 🙏 What video content would YOU point this at first? Drop it below 👇 #OpenRAG #RAG #Docling #Whisper #AI #SemanticSearch #VideoTranscription #OpenSource #AI #LLM #ContentCreators #Ollama

Modern #RAG isn't just about retrieving #chunks anymore. It's about using #agentic systems that reason through what's needed, retrieve iteratively, and validate context before they answer. It's also not about blowing all of your #tokens ingesting documents over and over. Link in reply.

Hey friends. There's a lot happening in the world of "context" these days. Huge #context windows and #models that are more and more capable. Does this mean #RAG is finally dead? I address this in my latest video comparing RAG and direct context. Check it out. Link in reply.

As a #developer and content creator I have to scan through LOTS of videos on a regular basis. Now I don't need to. I learned the other day that Docling is not only for ingesting and parsing documents. It turns out it can manage audio and video as well and easily pass that data into a #RAG pipeline. Now I just ask questions and the #agent tells me where to look. I made a video about it. Link in the reply. #openrag #docling