No, these are not about GPT-4.5, Claude 3.7 Sonnet, and Grok-3.
1. Mastra is a Typescript framework that gives you everything to build AI agents, RAG, automated workflows, and agent ops. Run it locally or deploy to a serverless cloud. 100% opensource.
2. AI agents fail in production -- agents crashing, message coordination failing, and expensive LLM calls. Orra is a plan engine designed specifically for multi-agent AI applications that handles the complex orchestration between your AI agents and execution environment.
3. Replit released Replit Agent v2, a more autonomous AI coding agent powered by Claude 3.7 Sonnet, featuring an industry-first realtime app design preview that renders interfaces as they're being built. Available in early access to paid subscribers who opt for Explorer Mode.
4. GenSX is a Typescript framework for building agents and workflows with reusable React-like components. Instead of graph-based approaches with nodes, edges, and global state, it uses JSX to define the flow of data, providing a visually clear, top-to-bottom structure.
5. Cloudflare released an opensource AI Agent framework to build agents that persist state, execute tasks, browse the web, and call AI models in real-time. You can deploy theseAI agents directly on Cloudflare's Workers platform.
We have created 50+ AI Agents and RAG tutorials, 100% free and opensource.
5 breakthroughs you missed while everyone talked about OpenAI.
1. Cloudflare just dropped an AI Agent framework to build agents that persist state, execute tasks, browse the web, and call AI models in real-time. 100% opensource.
2. Inception, AI startup founded by a Stanford professor, released Mercury Coder—the first commercial diffusion-based language model that generates text 5-10x faster than current LLMs. It achieves over 1000 tokens/ second on standard hardware.
3. LangChain released LangGraph Swarm, a lightweight Python library built on top of LangGraph, to create swarm-style multi-agent systems where specialized AI agents collaborate by dynamically handing off control based on their expertise.
5 alternatives to OpenAI's $200/month Operator Agent.
100% free.
WebUI: Open-source framework to facilitate interaction of AI agents with web browsers through a user-friendly GUI using Gradio. Built on Browser-use framework, it lets you use your own browser and maintain browser sessions between AI tasks.
Stagehand: Open-source framework to create AI web browsing agents. It's built on top of Playwright, a popular browser automation library. With 3 simple APIs (act, extract, and observe), it provides building blocks for web automation via natural language.
AI Agents, RAG and LLM releases in the last 24 hours
Low-code AI agent framework with self-reflection
Run AI agents in isolated sandboxes
DeepSeek V3 Base MoE model open-sourced
Microsoft’s open-source tool for prompt optimization
Built-in GPU memory visualizer in PyTorch
PraisonAI combines existing agent frameworks like AutoGen and CrewAI into a flexible low-code and no-code solution that makes building multi-agent systems more approachable. It also has built-in self-reflection capabilities, real-time voice interaction, and tools integrated.
MarinaBox, a new open-source toolkit, provides isolated desktop and browser sandboxes via Docker containers. It also has live session viewing, recording capabilities, and human-in-the-loop interactions. Run everything locally with the same API calls you'll use in production.
Auto-RAG for intelligent RAG pipelines
Open-source framework to build stateful AI agents
Built-in code runner in ChatGPT Canvas
Amazon's R&D lab to build foundational AI agent capabilities
Here’s everything you need to know:
Auto-RAG is a new iterative retrieval model that gives LLMs autonomous decision-making capabilities for retrieving and using external knowledge. LLMs independently determine when and what information to retrieve through multi-turn dialogues with the retriever.
Letta is an open-source LLM framework to build AI agents with persistent memory and complex reasoning capabilities. Letta is model-agnostic so you can maintain full visibility and control over how your agents think and remember.