← CookbookView source on GitHub ↗

LangChain deepagents + Nemotron @ Nebius Token Factory

A starter example of a LangChain Deep Agent powered by an Nvidia Nemotron LLM served by Nebius Token Factory.

Deep agents extend a regular tool-calling agent with:

Setup

This project uses uv for dependency management.

cd agents/nemotron-agents
uv sync

Create a .env file from the template and fill in your API keys:

cp env.example .env

Edit .env:

# get key from https://tokenfactory.nebius.com/
NEBIUS_API_KEY=your-nebius-api-key

Run the Agent

file: esearch_agent_1_nemotron.py

Bare deep agent (no external tools):

uv run python research_agent_1_nemotron.py

The report is written to output.md, and a short summary is printed to stdout. You can view a sample output here.

Run the Agent with Metrics

file : research_agent_2_nemotron_metrics.py

This agent will print out metrics like - tool calls - tokens count ..etc

Run it

uv run python research_agent_2_nemotron_metrics.py

you will see output similar to

--- Run Summary ---
Call #   Tool calls   Input tokens   Output tokens
---------------------------------------------------
1        1            3,361          328
2        1            4,124          77
3        1            17,022         101
4        0            18,427         849
---------------------------------------------------

Total input tokens:  42,934
Total output tokens: 1,355
Total tokens:        44,289

References