> For clean Markdown content of this page, append .md to this URL. For the complete documentation index, see https://docs.nvidia.com/dynamo/llms.txt. For full content including API reference and SDK examples, see https://docs.nvidia.com/dynamo/llms-full.txt.

# AI Simulate (Experimental)

<Warning>
**Experimental.** AI Simulate is intended for evaluation and feedback, not production capacity
planning. Its Python APIs, configuration schemas, search results, and deployment output may
change without a standard deprecation period. It provides no SLA, accuracy, or
configuration-optimality guarantees.
</Warning>

AI Simulate is a standalone Python distribution in the Dynamo repository. It contains simulation
and configuration-search tools that use Dynamo models and replay without making those tools part
of the stable `ai-dynamo` Python API.

## Spica

[Spica](/dynamo/dev/knowledge-base/modular-components/ai-simulate/spica/overview) is AI Simulate's first package. It searches engine, router, Planner, and
KV Block Manager (KVBM) settings with a black-box optimizer, evaluates candidates with Dynamo
Replay, and supports scalar and Pareto-front objectives.

Spica requires the matching Dynamo runtime from the same source revision. Follow the
[Spica development setup](/dynamo/dev/knowledge-base/modular-components/ai-simulate/spica/overview#develop), then invoke the package module:

```bash
python -m aisimulate.spica --config examples/aisimulate/spica/configs/smart_sweep.yaml
```

The `dynamo-planner` container builds and installs this distribution. AI Simulate is not included
in the `ai-dynamo` wheel and does not add an `ai-dynamo[spica]` extra or a console script.