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

# Not Diamond ¬◇

> Use Not Diamond to route prompts to the right model for your needs while maximizing accuracy and saving on costs

When you build complex LLM workflows, you may need to prompt different models according to accuracy,
cost, or call latency. You can use [Not Diamond][nd] to route prompts in these workflows to the
right model for your needs, helping maximize accuracy while saving on model costs.

This guide shows you how to integrate Not Diamond with W\&B Weave so that Weave automatically traces
routed model calls, and how to train a custom router using Weave Evaluations to route prompts based
on your own performance data.

## Get started

To use Not Diamond with Weave, you must have a Not Diamond account and an API key. [Create an
account][account] and [generate an API key][keys], then add your API key to your environment as
`NOTDIAMOND_API_KEY`.

<img src="https://mintcdn.com/wb-21fd5541/S0cRiDzxeODX77LU/weave/guides/integrations/imgs/notdiamond/api-keys.png?fit=max&auto=format&n=S0cRiDzxeODX77LU&q=85&s=74a45e52c932eaff2094855c07831f0d" alt="Create an API key" width="3454" height="1912" data-path="weave/guides/integrations/imgs/notdiamond/api-keys.png" />

From here, you can do any of the following:

* Try the [quickstart guide].
* [Build a custom router][custom router] with Weave and Not Diamond.
* [Chat with Not Diamond][chat] to see routing in action.

## Tracing

Weave integrates with [Not Diamond's Python library][python] to [automatically log API calls][ops],
so you can inspect routing decisions and provider responses alongside the rest of your Weave traces.
Run `weave.init()` at the start of your workflow, then continue to use the routed
provider as usual:

```python lines theme={null}
from notdiamond import NotDiamond

import weave
weave.init('notdiamond-quickstart')

client = NotDiamond()
session_id, provider = client.chat.completions.model_select(
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Concisely explain merge sort."}
    ],
    model=['openai/gpt-4o', 'anthropic/claude-3-5-sonnet-20240620']
)

print("LLM called: ", provider.provider)  # openai, anthropic, etc
print("Provider model: ", provider.model) # gpt-4o, claude-3-5-sonnet-20240620, etc
```

## Custom routing

For specialized use cases, you can train your own [custom router] on [Evaluations][evals], which lets
Not Diamond route prompts according to eval performance rather than the default routing policy.

Start by training a custom router. The `train_router` call returns a `preference_id` that identifies
your trained router for later use:

```python lines theme={null}
from weave.flow.eval import EvaluationResults
from weave.integrations.notdiamond.custom_router import train_router

# Build an Evaluation on gpt-4o and Claude 3.5 Sonnet
evaluation = weave.Evaluation(...)
gpt_4o = weave.Model(...)
sonnet = weave.Model(...)

model_evals = {
    'openai/gpt-4o': evaluation.get_eval_results(gpt_4o),
    'anthropic/claude-3-5-sonnet-20240620': evaluation.get_eval_results(sonnet),
}
preference_id = train_router(
    model_evals=model_evals,
    prompt_column="prompt",
    response_column="actual",
    language="en",
    maximize=True,
)
```

When you pass this preference ID to any `model_select` request, you can reuse your custom router to
route prompts to maximize performance and minimize cost on your evaluation data:

```python lines theme={null}
from notdiamond import NotDiamond
client = NotDiamond()

import weave
weave.init('notdiamond-quickstart')

session_id, provider = client.chat.completions.model_select(
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Concisely explain merge sort."}
    ],
    model=['openai/gpt-4o', 'anthropic/claude-3-5-sonnet-20240620'],

    # passing this preference ID reuses your custom router
    preference_id=preference_id
)

print("LLM called: ", provider.provider)  # openai, anthropic, etc
print("Provider model: ", provider.model) # gpt-4o, claude-3-5-sonnet-20240620, etc
```

## Additional support

For further support, visit the [docs] or [send a message][support].

[account]: https://app.notdiamond.ai

[chat]: https://chat.notdiamond.ai

[custom router]: https://docs.notdiamond.ai/docs/router-training-quickstart

[docs]: https://docs.notdiamond.ai

[evals]: /weave/guides/core-types/evaluations

[keys]: https://app.notdiamond.ai/keys

[nd]: https://www.notdiamond.ai/

[ops]: /weave/guides/tracking/ops

[python]: https://github.com/Not-Diamond/notdiamond-python

[quickstart guide]: https://docs.notdiamond.ai/docs/quickstart-routing

[support]: mailto:support@notdiamond.ai