deep-agents

The article discusses the integration of the open-source Stagehand SDK with Browserbase to provide ready-made browser tools for AI agents in LangChain, enabling easy setup for browser interaction.

LangChain introduces a new feature in its Deep Agents framework where subagents can now fork the supervisor's full conversation context instead of starting from a blank window.

Caspar discusses how LangChain's managed deep agents can automatically set up Slack and derive icons from agent names, with inspiration from DitherAvatar.

LangChain Academy is hosting a live workshop on Deep Agents on September 17th, teaching participants to build agents with planning, memory, and subagents for complex tasks.

Box adopted LangChain's Deep Agents framework to build its enterprise AI agent, highlighting benefits like model flexibility and faster development cycles for complex content management tasks.

LangChain introduces Managed Deep Agents (MDA), a simplified way to build and deploy production AI agents with managed infrastructure and a deep agent harness.

LangChain tested NVIDIA's open-source router Switchyard with Deep Agents, showing that routing between models can cut costs by ~70% while retaining ~90% accuracy, and introduced new middleware integration for NVIDIA models.

Austin Berke, Lead AI Product Engineer at Harmonic, shared how they used LangChain's Deep Agents framework to rebuild Scout, quadrupling user retention from week 1 to week 4. The core approach is to use the file system as shared storage to manage context, connecting agents, frontend, and deterministic processes, allowing the model to build context on demand.

LangChain shares a guide by Sydney Runkle explaining the differences and use cases for Deep Agents, LangChain, and LangGraph, with examples to help developers choose the right package.

LangChain Academy announces an update to its free 'Introduction to Deep Agents' course, adding modules on dynamic and async subagents, a course tutor skill, capstone projects, practice exercises, and TypeScript support, along with Deep Agents v0.7 updates.

LangChain releases Deep Agents v0.7, a leaner agent harness that cuts base input tokens by 65% through removing the default system prompt, trimming tool descriptions, and making TodoListMiddleware opt-in, with comparable performance validated across multiple models.

Tavily announces native integration with NVIDIA's NemoClaw Deep Agents Blueprint, enabling easy configuration for tool use and network policy governance.

LangChain shows a 7-minute tutorial by Partner Engineer Srimanth Tangedipalli on running Deep Agents Code inside a governed NemoClaw OpenShell Sandbox with NVIDIA Nemotron 3 Ultra via Baseten.

NVIDIA Nemotron 3 Ultra achieves benchmark-leading performance with LangChain Deep Agents harness, offering higher accuracy at lower cost than closed models without retraining.

LangChain Academy launches a new course titled 'Introduction to Deep Agents', teaching what a harness is and the four core capabilities of a harness, with hands-on building using LangChain and LangSmith.

LangChain announces support for recursive language model (RLM) workflows in Deep Agents, with a 6-minute explainer video.

LangChain has released a unified evaluation stack (Harbor + LangSmith) and an in-process code execution solution based on WASM+QuickJS, aimed at improving the evaluation reliability and execution safety of AI agents.

LangChain's Deep Agents allow untrusted agent-written code to run securely using WebAssembly-based code interpreters, providing execution and capability isolation without a traditional sandbox.

Harbor Framework now integrates with LangChain's Deep Agents, LangSmith Sandboxes, and Observability, enabling running agents in isolated, reproducible environments for deterministic testing.

LangChain's Deep Agents now supports dynamic subagents, where the main agent writes orchestration code to coordinate work at scale, enabling workflows like processing hundreds of documents with deterministic coverage.