Uploaded July 2026 | Updated November 2026, 1 day ago
In this video, Alex Olson from the LangChain team sets up dcode (LangChain's open source, model agnostic coding agent) with Nemotron 3 Ultra served through Baseten, then put it to work on a real coding task. Along the way we cover installation, model selection, authenticating LangSmith for full tracing, and the new /goal feature, before closing with a look at how NVIDIA's NemoClaw Deep Agents Blueprint takes this same setup into a secure, governed environment for enterprise use.
Chapters:
0:00 A new era for open models in agent engineering
0:11 Meet Nemotron 3 Ultra
0:47 Why the harness matters as much as the model
1:07 Introducing dcode, a provider agnostic coding agent
1:39 Installing dcode
2:08 Selecting Nemotron 3 Ultra as your model
2:38 Connecting your Baseten API key
3:04 Kicking off a long running coding task
3:39 Authenticating LangSmith with /auth
4:07 Exploring threads, offload, and MCP
4:27 Trying the new /goal feature
5:15 Checking in on the finished chat app
5:25 Inside LangSmith, tracing your agent
6:42 Taking dcode to production with NemoClaw
Resources:
→ Deep Agents Code (dcode): langchain.com/deep-agents
→ LangSmith: langchain.com/langsmith
→ LangChain Academy: academy.langchain.com
In this video, Alex Olson from the LangChain team sets up dcode (LangChain's open source, model agnostic coding agent) with Nemotron 3 Ultra served through Baseten, then put it to work on a real coding task. Along the way we cover installation, model selection, authenticating LangSmith for full tracing, and the new /goal feature, before closing with a look at how NVIDIA's NemoClaw Deep Agents Blueprint takes this same setup into a secure, governed environment for enterprise use.
Chapters:
0:00 A new era for open models in agent engineering
0:11 Meet Nemotron 3 Ultra
0:47 Why the harness matters as much as the model
1:07 Introducing dcode, a provider agnostic coding agent
1:39 Installing dcode
2:08 Selecting Nemotron 3 Ultra as your model
2:38 Connecting your Baseten API key
3:04 Kicking off a long running coding task
3:39 Authenticating LangSmith with /auth
4:07 Exploring threads, offload, and MCP
4:27 Trying the new /goal feature
5:15 Checking in on the finished chat app
5:25 Inside LangSmith, tracing your agent
6:42 Taking dcode to production with NemoClaw
Resources:
→ Deep Agents Code (dcode): langchain.com/deep-agents
→ LangSmith: langchain.com/langsmith
→ LangChain Academy: academy.langchain.com










