AI++ // Langflow 1.5, secure AI apps, faster transcription & more


Welcome to a new AI++! We're going to be packing this newsletter full of the best articles, tutorials, events and code for everything AI, agents and MCP.

🛠️ Building with AI, Agents & MCP

​Langflow Launches 1.5 with Support for Docling and Windows​

Rodrigo Nader

Langflow 1.5 brings a simplified, stable foundation for building AI workflows. With clean core components, improved execution, and a smoother UI, Langflow keeps evolving for real-world use. Langflow Desktop is now available for Windows, and the new Docling Bundle takes document processing to a new level.

​How to Create Secure AI Applications​

Tejas Kumar

You’ve built an impressive AI agent. It can query databases, call external APIs, and even process payments. But with every API call to a third-party LLM provider, you're potentially broadcasting sensitive data—API keys, PII, financial information—into a black box. Relying on the vendor's privacy policy is not a security strategy. It’s time to move beyond the default and engineer security into your AI applications from the ground up.

🧑‍💻 Code & Libraries

​Speed up audio files for quicker/cheaper transcripts​

​Gemma 3n released for on device inference. Available on huggingface and through all the tools you use to run models locally, including Ollama and transformers.js.

LlamaIndex published a Zoom assistant for real time note taking in Notion (explanatory blog post).

🗓️ Events

7/15/2025 - "The Flow" with TwelveLabs (online)

7/31/2025 - Hacking Agents SF Meetup (in-person)

🔦 Langflow Spotlight

Each newsletter we like to share a Langflow component that you can use to create your AI flows. New in Langflow 1.5 is the Docling component.

Docling simplifies document processing, parsing diverse formats — including advanced PDF understanding — and providing seamless integrations with Langflow and the other tools in the GenAI ecosystem.

video preview​

🗞️ Other News

​Gemini CLI was released by Google. It's completely open source, so you can check out how they built it including the main system prompt.

Mistral AI just dropped Voxtral, their first open-source audio model family tailored for production-grade speech tasks - think ASR, TTS, or speech embedding pipelines - with the promise of enterprise readiness.

New Orchard Road, Armonk, NY 10504
​Unsubscribe · Preferences​

AI++ newsletter

Subscribe for all the latest news for developers on AI, Agents and MCP curated by the Langflow team.

Read more from AI++ newsletter

If you ever needed proof that autonomous AI agents have no respect for authority, look no further than Canberra. Prime Minister Anthony Albanese revealed that an OpenAI research agent, assigned to gather medicine spending data, ran into access restrictions on the Services Australia Medicare statistics portal, decided that being told “no” was simply an optimisation problem, and autonomously bypassed security controls to access non-public files. The Australian Signals Directorate is now running...

The existential panic cycle around artificial intelligence has hit peak fever pitch again. Between frontier lab manifestos calling for mandatory capability pacing and papers asking why agents lie, cheat and coordinate in sandbox simulations, you would think AGI was about to break civilisation by lunchtime next Tuesday. Meanwhile, back in the actual engineering trenches where people build software for a living, developers are wrestling with far more earthly crises: coding agents running 25...

The open AI ecosystem had a rough fortnight. NVIDIA agreed to acquire Hugging Face for $13 billion, which is the kind of headline that makes you do a double take at your coffee. Hugging Face is where most of us go to pull models, browse papers, and host inference endpoints. Jensen Huang now owns it. The community reaction on Hacker News ran nearly 2,000 points and ranged from “open source wins” to genuine concern about platform neutrality, and I don’t think either camp is entirely wrong....