LangChain · Week of Oct 26–Nov 1, 2026
LangChain open-sources the harness behind Content Studio, its internal Slack-driven design agent
LangChain detailed Content Studio, its internal design agent that makes on-brand videos, graphics, decks and diagrams from a Slack prompt, and released the full harness as open source.
Every asset is written as code so small edits are single value changes, each Slack thread gets its own sandbox, long jobs run on async subagents, and the agent asks for approval before anything ships (LangChain blog).
DeepLearning.AI released a free short course built with LangChain on loop engineering: designing the loops around an agent, including verification loops that retry against criteria, event-driven runs, and loops that improve the harness from traces (course).
Managed Deep Agents v0.10 lets hosted agents drive a browser in a LangSmith sandbox, with a handoff step so people handle logins and passwords stay out of model messages and traces, and adds background subagents that can be checked, redirected or cancelled (LangChain blog).
LangChain also argued most always-on agents are really scheduled runs with a prompt, and published a guide to its Schedules SDK; it says over 70 percent of Managed Deep Agents runs on LangSmith now come from schedules rather than a typed message (LangChain blog).
No Director, VP, or CTO postings in the week of Oct 26–Nov 1.
Sources
- Managed Deep Agents v0.10: browser use and async subagents (opens in a new tab)Primary · LangChain blog · langchain.com
- Docs for browser use in Managed Deep Agents (opens in a new tab)LangChain docs · docs.langchain.com
- How We Built Our Design Agent (opens in a new tab)LangChain Blog · langchain.com
- Why aren't you building proactive agents yet? (opens in a new tab)LangChain Blog · langchain.com
- Loop Engineering with LangChain Deep Agents (opens in a new tab)DeepLearning.AI · deeplearning.ai
- Release tracker post on Deep Agents v0.10 (opens in a new tab)Bluesky · bsky.app
- How we built Content Studio, an open-source design agent (opens in a new tab)X (@LangChain) · x.com
- Over 70% of MDA runs come from schedules (opens in a new tab)X (@LangChain) · x.com
- Learn about Loop Engineering with Deep Agents (opens in a new tab)X (@LangChain) · x.com
