LangChain open-sources design-agent, the Slack design agent it runs internally as Content Studio
design-agent is LangChain's open-source (Apache 2.0) design agent, the one it runs internally as Content Studio. Built on Deep Agents and deployed on Managed Deep Agents, it takes requests in Slack, the API or a terminal and returns on-brand stills, decks, diagrams, charts and motion graphics.
- First seen
- (via GitHub new repos: topic ai-agent)
- Source published
- (we saw it 10.4 days later)
- Brief published
What changed
- LangChain published design-agent under the Apache 2.0 license; the README says it is the same design agent the company runs internally under the name Content Studio.
- It is a Deep Agents agent deployed on Managed Deep Agents (public beta on LangSmith Cloud), which supplies the Slack app, sandboxes, memory, schedules and tracing.
- Each Slack thread gets its own LangSmith sandbox with Chrome, FFmpeg, HyperFrames, D2 and fonts; slides are HTML and films are animation timelines, so edits re-render instead of restarting.
- A router skill picks one of 35 skills per request, and code checks colours, fonts, contrast, sizes and chart math against the brand pack on every render.
- Claude Opus 5.5 is the default model; OpenAI and Google models, OpenAI-compatible endpoints and the LangSmith Gateway are listed as supported, and a Docker self-host path exists.
- The README reports that since August it has shipped over 3,000 assets at LangChain, and that a keyless local demo runs with Docker and a scripted model.
In the source’s words
“Content Studio is the design agent LangChain runs, now open source.”
“at LangChain the median request costs about 50 cents in model spend, and long films and decks cost more.”
“The sandbox never holds a credential and has no network by default.”
What we have not verified
- We did not install, deploy or run design-agent ourselves.
- The usage figures (3,000+ assets, about 4.5M views, three in four requests from outside marketing) and the 50-cent median cost are LangChain's own claims.
- Managed Deep Agents hosting pricing is not stated in the README.
- The source-published time is the repository creation time on GitHub, not a dated announcement.
This brief is Reported: written from the sources below, not from running the release ourselves.
Sources
- langchain-ai/design-agent on GitHub (README) (primary source) (fetched 2026-10-28)
Related
Spotted an error? See our corrections policy or email the address on the contact page. Markdown version for agents: /news/langchain-open-sources-design-agent.md
Want these briefs in your Google Top Stories? Add Agentic Weekly as a preferred source on Google.
Free · weekly · about six minutes
Get The Weekly Verdict, free
What shipped this week, what is worth switching to, and what to ignore. We run what we can and label what we only read.
- One email a week. Each release gets one call: update now, worth a look, or skip.
- Every claim shows its evidence. Reported means we read the sources; Tested means we ran it, with the exact scope and the failures left in.
- A free AI Tools Cheat Sheet (web page and printable PDF) lands in your inbox when you confirm.
Free. Unsubscribe anytime. No spam. We email a confirmation link first and add you only after you click it. We use your email only to send this newsletter and never sell it. Privacy