> ## Documentation Index
> Fetch the complete documentation index at: https://docs.langchain.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Fine-tune models with Smithtune

> Use Smithtune to fine-tune models on LangSmith conversations and compare the results in LangSmith.

<Note>
  Smithtune is in [beta](/langsmith/release-stages).
</Note>

Smithtune is a command-line tool for fine-tuning models on [trajectories](/langsmith/observability-concepts#trajectories) recorded in LangSmith. It trains a model with [Fireworks](https://fireworks.ai/) or [Baseten](https://www.baseten.co/) and compares the base and tuned models in a LangSmith [experiment](/langsmith/evaluation-concepts#experiment). You can optionally deploy the tuned model as an endpoint for your application.

## How Smithtune works

<Steps>
  <Step title="Create a dataset">
    Select trajectories from a tracing project with a [trace query](/langsmith/trace-query-syntax) filter, a model judge, or both.
  </Step>

  <Step title="Prepare data">
    Validate the trajectories, then split them into training, validation, and test data. Each trajectory stays in one split.
  </Step>

  <Step title="Train a model">
    Preview the run, then fine-tune a supported model with your provider.
  </Step>

  <Step title="Compare results">
    Evaluate the base and tuned models on the test data, then [compare the experiments](/langsmith/compare-experiment-results) in LangSmith. Evaluation does not require a deployed endpoint.
  </Step>

  <Step title="Deploy an endpoint (optional)">
    Serve the tuned model for your application. To run an agent that uses it in production, see [LangSmith Deployment](/langsmith/deployment).
  </Step>
</Steps>

<Tip>
  To have your coding agent drive the fine-tuning process end to end, use the [Smithtune skill](https://github.com/langchain-ai/smithtune/blob/main/src/smithtune/skills/smithtune/SKILL.md).
</Tip>

Evaluation scores measure how closely each model matches the recorded behavior, not whether it completes tasks end to end. Smithtune does not execute the tool calls the model generates.

## Set up Smithtune

Smithtune requires trajectories in a tracing project or a LangSmith trajectory dataset. You also need API keys for LangSmith, your training provider, and a judge model. Training, evaluation, and deployed endpoints incur charges from the services you use.

To set up Smithtune:

1. Install Smithtune and the [LangSmith CLI](/langsmith/langsmith-cli), and set your API keys. Follow the [Smithtune README](https://github.com/langchain-ai/smithtune#readme) for the install command and the environment variables each provider needs.

2. Read the [data rights and permitted use terms](https://github.com/langchain-ai/smithtune/blob/main/docs/data-rights-and-permitted-use.md) linked from the [README](https://github.com/langchain-ai/smithtune#readme), then acknowledge them before your first run:

   ```bash theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
   smithtune acknowledge-data-rights
   ```

3. Check your local setup:

   ```bash theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
   smithtune doctor
   ```

For the commands in each step, see the [README](https://github.com/langchain-ai/smithtune#readme) or run `smithtune --help`.

## See also

* [Trajectory evaluations](/langsmith/trajectory-evals)
* [Query threads](/langsmith/query-threads)
* [Manage datasets](/langsmith/manage-datasets)

***

<div className="source-links">
  <Callout icon="terminal-2">
    [Connect these docs](/use-these-docs) to your agent of choice via MCP for real-time answers.
  </Callout>

  <Callout icon="edit">
    [Edit this page on GitHub](https://github.com/langchain-ai/docs/edit/main/src/langsmith/smithtune.mdx) or [file an issue](https://github.com/langchain-ai/docs/issues/new/choose).
  </Callout>
</div>