IBM Developer

🚀 Exciting Collaboration Announcement at CES: NVIDIA × Docling 🚀 We’re thrilled to share that NVIDIA and Docling are collaborating to deliver best-in-class performance for high-quality document conversion. This collaboration has already led to significant optimizations in Docling, leveraging batch processing to achieve up to a 4× speed-up compared to CPU-only workloads. ️Thanks to the close partnership with the NVIDIA RTX AI team, we were able to develop, test, and benchmark across multiple NVIDIA systems, including the powerful RTX 5090—pushing the limits of document conversion performance on RTX PCs. Learn more about the NVIDIA RTX optimizations: https://lnkd.in/dpsHq7af Get started with Docling on RTX: https://lnkd.in/dsN4KHwt We’re excited about what’s ahead and look forward to enabling faster, higher-quality document processing for the community. Stay tuned for more Docling news! Many thanks to our awesome collaborators at Nvidia: Aslı Sabancı Demiröz, Douglas O'Flaherty, Sean Sodha and many more! cc: Abdel Labbi, Sriram Raghavan, Shankar Ramaswamy, Ed Anuff, Milan (Marquez) Higginbotham, Alejandro Cantarero #CES #IBM #NVIDIA #Docling #RTXAI #OpenSource #DocumentAI #Performance #AIInfrastructure

https://lnkd.in/efqRppzd Docling + Go + Bob: The Modern Document Stack

One last look before logging off for the holidays 👀 // nothing on fire // see you 2026-0105T00:00:00Z

Langflow 1.7 is here, just in time for hacking over the holidays! We've got a ton of goodies for developers to unwrap 🎁 Here are the highlights 👇 🌐 Streamable HTTP for MCP (clients + servers) Langflow now supports Streamable HTTP across both MCP clients and servers. That means you can connect to any MCP server regardless of transport—and expose your Langflow projects the same way. 🤖 New agent components: ALTK & CUGA • ALTK improves agent reliability with tool validation and smart JSON post-processing (no more bloated context windows). • CUGA (from IBM Research) enables agents that can plan, break down complex tasks, write code, browse the web, and act across tools. These are serious upgrades for real-world agent workflows. 🔐 Authenticated webhooks Webhook triggers now support API key authentication—making it much safer to run Langflow flows in production. ☁️ AWS S3 file storage Uploaded files can now be stored in S3 instead of local disk, making Langflow a much better fit for cloud and team deployments. 🧩 A wave of new components LLM Selector, Smart Router, CometAPI (500+ models), AWS Bedrock Converse, flow control, mock data, and more. 👉 Check out the release, star the repo, join the community—and start building. Can’t wait to see what people ship with this 🚀 https://lnkd.in/gQDa7YkK

I just got early access to IBM Bob, and I was so excited to try it that I lost sleep, it’s 5:00 AM now. 😴 In just a few minutes, with no manual work beyond giving Bob the approval. Bob helped me build IBM watsonx Orchestrate agent that consumes MCP tools, after quickly building, testing, deploying, and validating the MCP server end to end: ⦿ Build MCP server with two tools ⦿ Generate documentation ⦿ Create unit tests and run them ⦿ Deploy the MCP server, plus an MCP client to validate it end to end ⦿ Import the MCP tools into watsonx Orchestrate ⦿ Create a watsonx Orchestrate agent that consumes those MCP tools The possibilities and productivity gains are endless! Stay tuned on IBM Bob tutorials soon! Meanwhile, you can see some screenshots in the comments.

As 2025 comes to a close, we’re proud of the momentum behind Docling, an LF AI & Data project focused on making documents truly ready for GenAI. This year, Docling continued to grow as an open source framework for: 🔹 Parsing complex documents like PDFs, DOCX, PPTX, images, and more 🔹 Preserving layout, tables, charts, formulas, and reading order 🔹 Supporting secure, local execution for enterprise and regulated environments Thank you to the Docling maintainers and contributors for pushing document AI forward. More to come in 2026. 👉 Get involved with Docling: https://hubs.la/Q03YF1RW0

Markus Eisele shares three ways Quarkus modernizes Java, based on how teams actually build and ship today. Swipe to explore ↓

ClickOps doesn’t scale. John Wen breaks down how a Python SDK enables data teams to build pipelines in code—making them easier to develop, version, automate, and govern to power agents with AI-ready data. ↓

🚀 ContextForge MCP Gateway v1.0.0-BETA-1 Released Excited to share the first 1.0.0 beta milestone for ContextForge MCP Gateway - an open-source gateway for managing and securing MCP servers for AI agents. Key highlights: 🖥️ NEW: ContextForge Desktop App & CLI now available. 🏗️ Multi-Architecture Support - Now available for ARM64 (Apple Silicon, AWS Graviton) and IBM Z (s390x). ⚡ 10x+ Performance Gains - Concurrent health checks and N+1 query elimination dramatically reduce latency when hundreds of MCP servers are connected. Rust rewrite for PII Plugin. Additional Rust components. 🌐 Air-Gapped Deployments - All frontend assets bundled for fully offline environments - ideal for government, financial, and industrial use cases. 🔌 gRPC Translation (Experimental) - New high-performance gRPC interface for MCP operations, perfect for polyglot microservice environments. 🔐 Enterprise Security - Password expiration policies, one-time authentication, and gateway-level input validation for path traversal and injection protection. 🎉 Ecosystem Growth - Introducing ContextForge Desktop (native GUI) and ContextForge CLI for scriptable operations, hosted under the new contextforge-org GitHub organization. ✅ 25+ issues resolved in this release. 👉 Article: https://lnkd.in/gWGuJ8vN 🔗 GitHub: https://lnkd.in/eXv3fVqQ 📚 Docs: https://lnkd.in/eR5JmE4v Apache 2.0 licensed. Contributions welcome! Leave us a ⭐ on GitHub! #MCP #AI #OpenSource #AIAgents #DevOps #IBM #MachineLearning #ArtificialIntelligence