Milvus, created by Zilliz

Here's a full tutorial about how to use Claude code free with Ollama! You need to watch this! ——— 👉 Follow Milvus, created by Zilliz, for everything related to unstructured data!

𝐂𝐥𝐚𝐰𝐝𝐛𝐨𝐭 🦞 — 𝐭𝐡𝐞 𝐡𝐨𝐭𝐭𝐞𝐬𝐭 𝐀𝐈 𝐩𝐫𝐨𝐣𝐞𝐜𝐭 𝐫𝐢𝐠𝐡𝐭 𝐧𝐨𝐰 GitHub stars shot from 5k to 95k in days. Mac mini sold out because of it. What is it? An open-source personal assistant that lives in WhatsApp, Telegram, iMessage — wherever you already chat. Runs locally 24/7. The promise: "AI finds you" instead of "you find AI." But it's still early — setup requires CLI skills, security gaps have been identified, and some users reported the bot taking unintended actions. One thing worth noting: Clawdbot stores core memory in local Markdown files — simple but limited. As personal agents need to recall months of context, vector databases like Milvus become essential. Clawdbot's always-on presence + Milvus-powered semantic memory = AI that truly knows you. ——— 👉 Follow Milvus, created by Zilliz, for everything related to unstructured data!

🔧 𝗠𝗶𝗹𝘃𝘂𝘀 + Voyage AI by MongoDB integration guide Quality embeddings + millisecond vector search. A few lines of code. What you get:  → VoyageAI's state-of-the-art embedding models  → Milvus's lightning-fast vector retrieval  → Simple Python setup, production-ready 🚀 And if you haven't checked out the new Voyage 4 series yet — it's worth a look. Four models, each built for different needs:  → 𝘃𝗼𝘆𝗮𝗴𝗲-𝟰: solid all-rounder, balanced accuracy, cost & latency  → 𝘃𝗼𝘆𝗮𝗴𝗲-𝟰-𝗹𝗮𝗿𝗴𝗲: flagship model, best retrieval precision  → 𝘃𝗼𝘆𝗮𝗴𝗲-𝟰-𝗹𝗶𝘁𝗲: optimized for speed and cost  → 𝘃𝗼𝘆𝗮𝗴𝗲-𝟰-𝗻𝗮𝗻𝗼: open weights, great for local dev or on-device Whether you're building RAG pipelines, semantic search, or AI-powered recommendations — this combo just works. Full tutorial with code examples 👇 https://lnkd.in/gpH3Spbh ——— 👉 Follow Milvus, created by Zilliz, for everything related to unstructured data!

Running Large Collection 𝐌𝐢𝐥𝐯𝐮𝐬 𝐨𝐧 𝐊𝐮𝐛𝐞𝐫𝐧𝐞𝐭𝐞𝐬 𝐚𝐧𝐝 𝐡𝐢𝐭𝐭𝐢𝐧𝐠 𝐬𝐭𝐚𝐛𝐢𝐥𝐢𝐭𝐲 𝐢𝐬𝐬𝐮𝐞𝐬? If you’re 𝐬𝐜𝐚𝐥𝐢𝐧𝐠 𝐜𝐨𝐥𝐥𝐞𝐜𝐭𝐢𝐨𝐧𝐬 — especially in multi-tenant setups — here are the best practices Xiaofan(James) Luan, VP of Engineering at Zilliz, shared in our latest Milvus 2.6 webinar. If this doesn’t solve your problem, feel free to open an issue on the Milvus GitHub or reach out to us on Discord. Full webinar replay: https://lnkd.in/gFZeJqyE

Are 𝐊𝐚𝐟𝐤𝐚 & 𝐏𝐮𝐥𝐬𝐚𝐫 “𝐝𝐞𝐩𝐫𝐞𝐜𝐚𝐭𝐞𝐝”? And should we migrate to 𝐖𝐨𝐨𝐝𝐩𝐞𝐜𝐤𝐞𝐫 𝐢𝐧 𝐌𝐢𝐥𝐯𝐮𝐬 𝟐.𝟔? Here’s the takeaway from our 𝐌𝐢𝐥𝐯𝐮𝐬 𝟐.𝟔 𝐰𝐞𝐛𝐢𝐧𝐚𝐫: • Kafka & Pulsar are 𝐍𝐎𝐓 deprecated — for people who are already using Pulsar and Kafka, there's no need for migration yet • Woodpecker trades lower latency for easier maintenance and higher throughput, while Kafka/Pulsar add extra infrastructure + operational overhead • Choose based on workload: Pulsar for low latency, Woodpecker for batch ingestion 🎬 Watch the 1-min QA clip for full context 🎥 Full webinar replay: https://lnkd.in/gFZeJqyE

"𝐂𝐡𝐮𝐧𝐤 𝐟𝐢𝐫𝐬𝐭, 𝐞𝐦𝐛𝐞𝐝 𝐥𝐚𝐭𝐞𝐫" 𝐢𝐬 𝐰𝐫𝐨𝐧𝐠. 𝗠𝗮𝘅-𝗠𝗶𝗻 𝗦𝗲𝗺𝗮𝗻𝘁𝗶𝗰 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴 𝗳𝗹𝗶𝗽𝘀 𝘁𝗵𝗲 𝘀𝗰𝗿𝗶𝗽𝘁: 𝗲𝗺𝗯𝗲𝗱 𝗳𝗶𝗿𝘀𝘁, 𝗰𝗵𝘂𝗻𝗸 𝘀𝗲𝗰𝗼𝗻𝗱.  Instead of blindly splitting text by token count, it embeds all sentences upfront, then uses semantic similarity to decide where boundaries should go. 𝐓𝐡𝐞 𝐥𝐨𝐠𝐢𝐜 𝐢𝐬 𝐬𝐢𝐦𝐩𝐥𝐞:  If a new sentence is semantically close enough to the current chunk, it stays. If not, start a new chunk. Boundaries follow meaning, not arbitrary limits. It's not perfect (long-range context can still get fragmented), but for docs like API specs or release notes, it's a noticeable upgrade with minimal overhead. Learn more: https://lnkd.in/gC6cGRUC ——— 👉 Follow Milvus, created by Zilliz, for everything related to unstructured data!

Claude Code is now officially free to use. Ollama has supported Anthropic's Messages API, you can now run it 100% locally with open-source models. Get it running with this simple 5-step guide: [𝟭] 𝗜𝗻𝘀𝘁𝗮𝗹𝗹 𝗢𝗹𝗹𝗮𝗺𝗮 [𝟮] 𝗣𝘂𝗹𝗹 𝗮 𝘀𝘁𝗿𝗼𝗻𝗴 𝗼𝗽𝗲𝗻-𝘀𝗼𝗿𝘂𝗰𝗲 𝗰𝗼𝗱𝗶𝗻𝗴 𝗺𝗼𝗱𝗲𝗹 ollama pull qwen2.5-coder [𝟯] 𝗜𝗻𝘀𝘁𝗮𝗹𝗹 𝗖𝗹𝗮𝘂𝗱𝗲 𝗖𝗼𝗱𝗲 macOS, Linux, WSL: curl -fsSL https://lnkd.in/gERbb3ZM | bash Windows PowerShell: irm https://lnkd.in/gcuHBYSF | iex [𝟰] 𝗖𝗼𝗻𝗻𝗲𝗰𝘁 Point Claude Code to your local server instead of Anthropic's cloud: export ANTHROPIC_AUTH_TOKEN=ollama export ANTHROPIC_BASE_URL=http://localhost:11434 [𝟱] 𝗥𝘂𝗻 𝗜𝘁 claude --model qwen2.5-coder 💡 Pro tip: Add RAG with Milvus to let your local AI search your entire codebase. The barrier between proprietary tools and open-source models just disappeared. 🕊️ Try it now! ——— 👉 Follow Milvus, created by Zilliz, for everything related to unstructured data!

Building agents? Your context window is precious real estate. Semantic Highlight isn't just for RAG — it's a 𝐜𝐨𝐧𝐭𝐞𝐱𝐭 𝐩𝐫𝐮𝐧𝐢𝐧𝐠 𝐩𝐨𝐰𝐞𝐫𝐡𝐨𝐮𝐬𝐞 for agent applications. Same model, different superpower: ✂️ Trim irrelevant context before each LLM call 💰 Cut token costs dramatically 🎯 Keep your agent focused on what matters We trained on 𝟓𝟎𝟎𝐌+ 𝐛𝐢𝐥𝐢𝐧𝐠𝐮𝐚𝐥 𝐬𝐚𝐦𝐩𝐥𝐞𝐬 using Qwen3 8B for annotation — with full reasoning traces for quality control. The training data is 𝐚𝐥𝐬𝐨 𝐨𝐩𝐞𝐧-𝐬𝐨𝐮𝐫𝐜𝐞𝐝 on HuggingFace for anyone who wants to build on top. Cloud inference coming soon to Zilliz Cloud for plug-and-play deployment. Download it from Hugging Face 🤗: https://lnkd.in/g4F-zxRT ——— 👉 Follow Milvus, created by Zilliz, for everything related to unstructured data!

Semantic highlighting is changing the RAG game. 𝐏𝐫𝐨𝐮𝐝 𝐭𝐨 𝐬𝐞𝐞 𝐨𝐮𝐫 0.6𝐁 𝐛𝐢𝐥𝐢𝐧𝐠𝐮𝐚𝐥 𝐦𝐨𝐝𝐞𝐥 𝐛𝐞𝐢𝐧𝐠 𝐢𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐞𝐝 𝐢𝐧𝐭𝐨 𝐕𝐞𝐫𝐛𝐚𝐭𝐢𝐦𝐑𝐀𝐆 to make structured, citation-ready outputs even easier. 🛠️

𝐒𝐢𝐧𝐜𝐞 𝐌𝐢𝐥𝐯𝐮𝐬 2.6 𝐝𝐫𝐨𝐩𝐩𝐞𝐝, we've been hearing a lot from the community — many of you are excited about the new StreamingNode architecture, Woodpecker, RaBitQ quantization, and more. But we've also heard your questions: ❓"Should I use Helm or Operator for upgrading?" ❓"How do I choose between Pulsar, Kafka, and Woodpecker? ❓ Can I switch during the upgrade?" ❓ "Will my existing configs still work?" ❓ "How many resources does StreamingNode need?" 𝐖𝐞 𝐡𝐞𝐚𝐫 𝐲𝐨𝐮 🙌 𝐒𝐨 𝐰𝐞 𝐩𝐮𝐭 𝐭𝐨𝐠𝐞𝐭𝐡𝐞𝐫 𝐭𝐡𝐢𝐬 𝐅𝐀𝐐 𝐭𝐨 𝐡𝐞𝐥𝐩 𝐦𝐚𝐤𝐞 𝐲𝐨𝐮𝐫 𝐮𝐩𝐠𝐫𝐚𝐝𝐞 𝐣𝐨𝐮𝐫𝐧𝐞𝐲 𝐬𝐦𝐨𝐨𝐭𝐡𝐞𝐫.  If you have more questions, feel free to reach out — we're always here to help! 👇 ——— 👉 Follow Milvus, created by Zilliz, for everything related to unstructured data!