Documentation Index
Fetch the complete documentation index at: /llms.txt
Use this file to discover all available pages before exploring further.
Run an agent that you trained on Oumi inside an agent harness, as the main model or as a subagent
| As the main model | As a subagent | |
|---|---|---|
| How your agent starts | You select it as the session’s model | Your usual model calls one tool, for example ask_stock_agent |
| What your fine-tuned model sees | The harness’s requests, limited to your tools and system prompt | Only the requests that your subagent sends: your tools, your system prompt, and its own history |
| Who runs your agent’s tool loop | The harness | Your subagent code, inside a harness extension |
| Use it when | The session does only the work that your agent was trained for | Your agent does one specialized job inside broader work, such as coding or research |
Was this page helpful?
