March 2026: LangChain Newsletter
It feels like spring has sprung here, and so has a new NVIDIA integration, ticket sales for Interrupt 2026, and announcing LangSmith Fleet (formerly Agent Builder).
It feels like spring has sprung here, and so has a new NVIDIA integration, ticket sales for Interrupt 2026, and announcing LangSmith Fleet (formerly Agent Builder).
Harrison is in NYC this month and hosting a workshop—sign up here. Here's your monthly roundup of updates from LangChain.
Product Updates
LangSmith
🦜 Your favorite AI assistant, Polly, is now generally available everywhere in LangSmith. Polly can take action like an engineer on your team. Learn more.
🛠️ Agent Builder is now LangSmith Fleet. Along with the new name, Fleet now includes agent identity, sharing, and permissions, so you can manage your agent fleet across the company securely. Learn more.
🤹♀️ Skills are now available in LangSmith Fleet. Skills equip agents across your team with knowledge for specialized tasks. Learn more.
🏖️ LangSmith Sandboxes are now in Private Preview. Our sandboxes give your agents locked-down, temporary environments where they can run code safely, with granular control over access and resources. Learn more.
🚢 We introduced LangGraph’s Deploy CLI, a new set of commands in the langgraph-cli packages so you can deploy agents to LangSmith Deployment right from your terminal in just one step. Learn more.
🔒 Give Enterprise admins precise control over resource access with Attribute-Based Access Control (ABAC). Layer tag-based allow/deny policies on top of your existing RBAC roles to control who can access which projects, datasets, and prompts through the API. Learn more.
✍️ Track every administrative action across your LangSmith organization with Audit Logs. Get a tamper-resistant record of who did what and when—across members, workspaces, datasets, deployments, and more—queryable via API. Learn more.
Open Source
📈 langgraph v1.1 was released, including type-safe streaming, type-safe invoke, Pydantic and dataclass coercion, and more. This version is fully backwards compatible. Learn more.
🤖 deepagents v0.5.0 alpha release went live, including async subagents, multi-modal support, backend changes, and Anthropic prompt caching improvements. Learn more.
🥷 We released our first set of skills in our OSS ecosystem! Learn more.
Interrupt 2026 - Join 1,000+ builders and Jensen Huang
Join us May 13-14 in San Francisco for two days of talks, workshops, and lessons from teams shipping in production. This year's lineup includes AI teams from Clay, Rippling, and Honeywell sharing what's working (and what isn't), hands-on workshops with LangChain product experts, and keynotes from Harrison Chase, Jensen Huang, and Andrew Ng on what's coming next.
Speak the Lang
The anatomy of an agent harnessA model by itself isn't an agent. It becomes one when you wrap it in a harness — the system prompts, tools, middleware, memory, skills, and subagent orchestration that turn raw intelligence into something that can actually do work.
This post breaks down every layer of that harness:
- How hooks and middleware intercept model and tool calls to modify behavior
- How skills get loaded progressively (only when the agent determines they're relevant)
- How subagents enable context isolation and parallel execution
- How memory persists conventions and preferences across sessions
Read the full post →
Open SWE: an open-source framework for internal coding agentsOver the past year, teams at Stripe, Ramp, and Coinbase each built internal coding agents, and independently landed on nearly identical architectures: isolated cloud sandboxes, curated toolsets, subagent orchestration, and tight integration with developer workflows. Open SWE captures that pattern in an open-source framework you can fork and deploy yourself.
Read the full post →
How we built LangChain's GTM agentWe built an agent that runs our own outbound process end-to-end. Early results: lead-to-qualified-opportunity conversion up 250%, and reps reclaiming 40 hours a month each.
Read the full post →
Upcoming Events
LangChain Hosted
🏗️ (April 16) New York // AI Agents Workshop with Harrison ChaseWe're hosting a half-day workshop in New York on April 16, led by Harrison. He’ll kick things off with a talk on the current state of AI agents, followed by technical deep dives on building, observing, and deploying Deep Agents to production. RSVP here.
🌈 (April 22-24) Las Vegas // Google Cloud NextCome find us at Booth #5006 in the Expo Hall. The engineering team will be running demos all week, showing the latest LangSmith updates. Harrison will also be on-site taking meetings throughout the conference. In the evening, we're co-hosting a Happy Hour with MongoDB and Confluent. RSVP here.
🌴 (April 9) Miami // Community Meetup: Evals 102Evals 102 picks up where the basics left off. Expect a no-fluff deep dive into advanced evaluation patterns for AI agents. Organized by Ambassador Andres Torres at The Lab Miami. RSVP here.
🇦🇷 (April 10) Buenos Aires // Community Meetup: The LangChain StackOur first-ever LangChain Meetup in Argentina. Join us for an evening of quick talks and open discussion on the current LangChain stack. Organized by Ambassador Lucas Petralli. RSVP here.
🤝 Customer & integration highlights
LangChain announces our enterprise agentic AI platform built with NVIDIA. We're partnering with NVIDIA to give enterprise teams a full-stack platform for building and running production agents. As part of this, we're also joining NVIDIA's Nemotron Coalition to help advance frontier open models. Read the full post →
Moda uses a multi-agent system built with Deep Agents and LangSmith to give non-designers (i.e. marketers, founders, salespeople) a Cursor-style AI sidebar that creates and iterates on professional presentations, social posts, and brochures directly on a fully editable 2D vector canvas. Read their story.
How can you follow along with the Lang Latest? Check out the LangChain blog, Changelog, and YouTube channel for more product and content updates.
Join our newsletter
Updates from the LangChain team and community
Enter your email
Processing your application...
Success! Please check your inbox and click the link to confirm your subscription.
Sorry, something went wrong. Please try again.
Sign in to highlight and annotate this article
Conversation starters
Daily AI Digest
Get the top 5 AI stories delivered to your inbox every morning.
More about
integrationagentlangchain
Cortex Code in Snowflake: How to Use It Without Burning Credits
Snowflake Cortex Code (CoCo) is like an AI assistant inside Snowsight (and CLI also). You can ask it to write SQL, create dbt models, explore data, help in ML work, and even do some admin tasks. But one thing people don’t realise early — this tool is powerful, but also costly if used wrongly. Bad prompts → more tokens → more credits → surprise bill. Prompt Engineering (this directly impacts cost) CoCo works on token consumption. what you type → counted 2. what it replies → counted If your prompt is vague → more tool calls → more cost. Example: Bad: Help me with my data Good: Create staging model for RAW.SALES.ORDERS with not_null on ORDER_ID Best Practices: Use full table names 2. Be clear about output 3. Keep prompts small 4. Provide business logic upfront 5. Use AGENTS.md for consistency
Best Form Backend for Job Applications and Event Registrations in 2026
If you're collecting job applications or event registrations online, you've probably hit the same wall. Either you're overpaying for a tool like Typeform or JotForm, or you're cobbling together a Google Form that looks unprofessional and gives you zero control over where your data goes. In this article, I'll walk through the best form backends for job applications and event registrations in 2026, covering price, features, file upload support, and which one is right for your use case. Why the Right Form Backend Matters for Applications and Registrations A contact form getting 10 submissions a month is simple. A job application form getting 500 submissions a month is a different problem entirely. You need: File uploads: Candidates submit resumes, cover letters, and portfolios. High submissio
Knowledge Map
Connected Articles — Knowledge Graph
This article is connected to other articles through shared AI topics and tags.
More in Products
Cortex Code in Snowflake: How to Use It Without Burning Credits
Snowflake Cortex Code (CoCo) is like an AI assistant inside Snowsight (and CLI also). You can ask it to write SQL, create dbt models, explore data, help in ML work, and even do some admin tasks. But one thing people don’t realise early — this tool is powerful, but also costly if used wrongly. Bad prompts → more tokens → more credits → surprise bill. Prompt Engineering (this directly impacts cost) CoCo works on token consumption. what you type → counted 2. what it replies → counted If your prompt is vague → more tool calls → more cost. Example: Bad: Help me with my data Good: Create staging model for RAW.SALES.ORDERS with not_null on ORDER_ID Best Practices: Use full table names 2. Be clear about output 3. Keep prompts small 4. Provide business logic upfront 5. Use AGENTS.md for consistency
The Stack Nobody Recommended
The most common question I got after publishing Part 1 was some variation of "why did you pick X instead of Y?" So this post is about that. Every major technology choice, what I actually considered, where I was right, and where I got lucky. I'll be upfront: some of these were informed decisions. Some were "I already know this tool, and I need to move fast." Both are valid, but they lead to different trade-offs down the line. The Backend: FastAPI I come from JavaScript and TypeScript. Years of React on the frontend, Express and Fastify on the backend. When I decided this project would be Python, because that's where the AI/ML ecosystem lives, I needed something that didn't feel foreign. FastAPI clicked immediately. The async/await model, the decorator-based routing, and type hints that actu
Discussion
Sign in to join the discussion
No comments yet — be the first to share your thoughts!