Visual Programming IDE for AI Agents

Today’s top AI Highlights:

1. Open-source visual IDE to build production-ready AI agents

2. Build your own AI Data Science Agents team in days, no PhD required

3. RAG apps that retrieve data from a Knowledge Graph with a simple API call

4. xAI’s Grok standalone iOS app is around the corner

5. The only fully programmable platform for building AI voice agents

& so much more!

Read time: 3 mins

AI Tutorials

Building powerful RAG applications has often meant trading off between model performance, cost, and speed. Today, we're changing that by using Cohere's newly released Command R7B model - their most efficient model that delivers top-tier performance in RAG, tool use, and agentic behavior while keeping API costs low and response times fast.

In this tutorial, we'll build a production-ready RAG agent that combines Command R7B's capabilities with Qdrant for vector storage, Langchain for RAG pipeline management, and LangGraph for orchestration. You'll create a system that not only answers questions from your documents but intelligently falls back to web search when needed.

Command R7B brings an impressive 128k context window and leads the HuggingFace Open LLM Leaderboard in its size class. What makes it particularly exciting for our RAG application is its native in-line citation capabilities and strong performance on enterprise RAG use-cases, all with just 7B parameters.

We share hands-on tutorials like this 2-3 times a week, designed to help you stay ahead in the world of AI. If you're serious about leveling up your AI skills and staying ahead of the curve, subscribe now and be the first to access our latest tutorials.

Don’t forget to share this newsletter on your social channels and tag Unwind AI (X, LinkedIn, Threads, Facebook) to support us!

Latest Developments

Open-Source Visual Programming Environment for AI Agents 🥸

Rivet is an open-source visual IDE for building AI agents with LLMs. The tool lets you design, debug, and deploy complex prompt chains through an intuitive visual interface, with everything stored as YAML files for easy version control.

While other visual tools focus on prototyping, Rivet is built for production use - you can execute your prompt graphs directly in your application and debug them remotely. It is not just for prototyping; it’s built for production-level applications with real-time debugging, collaboration tools, and an integrated testing suite.

Key Highlights:

1. Visual Development - Build complex LLM chains using a node-based editor that gives you full visibility into data flow and execution state. Each node's input/output is visible in real-time, making it easier to identify issues. Your graphs are stored as YAML files, so you can version control them and review changes using standard tools.

2. Rich Debugging Capabilities - Debug your graphs both locally and remotely through a WebSocket connection. Watch live streaming outputs from LLMs, inspect the state of each node, and track executions across multiple runs. Pause execution mid-flow to examine the current state or step through complex chains node by node.

3. First-Class Testing Support - The built-in Trivet testing framework lets you validate your graphs against expected outputs. Write test cases with custom validation logic, run multiple iterations to account for LLM variance, and integrate testing into your CI/CD pipeline through the Trivet library.

4. Framework-Agnostic Integration - Run Rivet graphs in any Node.js or TypeScript application through a clean API. The core is designed to be lightweight and extensible - you can add custom nodes for your specific needs and integrate with existing tools and services through plugins.

Writer RAG tool: build production-ready RAG apps in minutes

• Writer RAG Tool: build production-ready RAG apps in minutes with simple API calls.

• Knowledge Graph integration for intelligent data retrieval and AI-powered interactions.

• Streamlined full-stack platform eliminates complex setups for scalable, accurate AI workflows.

Learn more about our production ready RAG tooling here.

Build Your Own AI Data Science Agents Team 📈

We've been saying this: 2025 will be the year of AI agents. These AI agents, equipped with the right tools, will plan, act, and execute tasks much like humans, automating a significant portion of our routine work. This won’t be limited to a specific domain; it’ll span every area, from legal and medicine to design and recruitment.

Here’s an agent platform capitalizing on this future: causaLens to build and deploy AI-powered data science agents within days. The platform includes pre-built agents for tasks like data cleaning and model deployment, and lets you create custom agents tailored to your organization’s needs. These agents can be used by all, not just data scientists, and use causal AI, which you can integrate into workflows.

Key Highlights:

1. Custom AI Data Science Agents - Build and deploy specialized AI agents trained on your data and optimized for your specific business logic. Agents develop expertise over time using long-term memory and a proprietary data science knowledge base.

2. Automated Workflows - Automate data science workflows using pre-built agents for tasks like planning, cleaning, analysis, causal inference, application building, and SQL queries. This multi-agent system completes complex analytic tasks, with you controlling and validating the results.

3. Causal Reasoning - The platform uses causal AI (CAIA Agent), going beyond simple correlations to identify cause-and-effect relationships. Integrate these workflows directly into your systems.

4. Secure Deployment - Connect with existing enterprise data platforms, using SOC 2, HIPAA, and ISO 27001 compliance. Deploy the platform flexibly to fit your workflows. You control the automation with intuitive "@" mentions and can pause, edit or rerun workflows.

Quick Bites

IBM, Princeton, CMU, and UIUC released Bamba-9B, a hybrid Mamba2 model trained on open data, boasting 2.5x better throughput and 2x faster latency than standard transformers in vLLM. This model is available now for experimentation in transformers, vLLM, TRL, and llama.cpp, along with tuning, training, and pretraining recipes – plus, they've open-sourced their stateless data loader.

Full-stack genAI platform Writer has launched a new RAG tool to quickly build production-ready RAG applications. Instead of complex configurations, you can now retrieve data from a Knowledge Graph using a single API call. This tool is integrated within the Writer platform and allows you to create scalable, highly accurate AI workflows by simply passing the graph ID of your data to the RAG tool.

Elon Musk's xAI is beta-testing a standalone iOS app for its Grok chatbot. The app, currently live in select regions, offers features like real-time web access and image generation. A dedicated website, Grok.com, is also coming soon.

Tools of the Trade

1. voice_dev: A programmable platform for building voice AI agents where you can deploy custom logic as serverless functions and integrate with any AI provider (OpenAI, Anthropic, etc.) for speech-to-text, text-to-speech, and LLMs. It handles all the infrastructure for real-time voice communication with ~300ms latency.

2. Shortest: Testing framework to write end-to-end tests in plain English, which are then executed using Anthropic's Claude API and Playwright under the hood. It integrates with GitHub for 2FA support.

3. RAG Logger: Open-source logging tool designed specifically for RAG apps. It serves as a lightweight, open-source alternative to LangSmith, focusing on RAG-specific logging needs.

4. Awesome LLM Apps: Build awesome LLM apps with RAG, AI agents, and more to interact with data sources like GitHub, Gmail, PDFs, and YouTube videos, and automate complex work.

Hot Takes

1. If you are a software engineer who’s three years into your career: quit now. there is not a single job in CS anymore. it's over. this field won't exist in 1.5 years. ~
   null
   
   

2. If I can’t install it simply with pip, then I simply won’t install it. ~
   Bojan Tunguz

That’s all for today! See you tomorrow with more such AI-filled content.

Don’t forget to share this newsletter on your social channels and tag Unwind AI to support us!

Unwind AI - X | LinkedIn | Threads | Facebook

Awesome LLM Apps | Sponsor Us

PS: We curate this AI newsletter every day for FREE, your support is what keeps us going. If you find value in what you read, share it with at least one, two (or 20) of your friends 😉