A high-performance inference engine for AI models
---
title: "Chat"
---
import StepsPython from '/snippets/quick-start/python-steps.mdx';
import StepsSwift from '/snippets/quick-start/swift-steps.mdx';
import StepsTypeScript from '/snippets/quick-start/typescript-steps.mdx';
import StepsRust from '/snippets/quick-start/rust-steps.mdx';
import CodePython from '/snippets/generated/examples/python/chat.mdx';
import CodeSwift from '/snippets/generated/examples/swift/chat.mdx';
import CodeTypeScript from '/snippets/generated/examples/typescript/chat.mdx';
import CodeRust from '/snippets/generated/examples/rust/chat.mdx';
<Tip>
In this example, we will download a model and get a reply to a specific list of messages.
</Tip>
<Tabs>
<Tab title="Python">
<Steps>
<StepsPython/>
</Steps>
</Tab>
<Tab title="Swift">
<Steps>
<StepsSwift/>
</Steps>
</Tab>
<Tab title="TypeScript">
<Steps>
<StepsTypeScript/>
</Steps>
</Tab>
<Tab title="Rust">
<Steps>
<StepsRust/>
</Steps>
</Tab>
</Tabs>
<Info>
Once loaded, the same `ChatSession` can be reused for multiple requests until you drop it. Each model may consume a significant amount of RAM, so it's important to keep only one session loaded at a time. For iOS apps, we recommend adding the [Increased Memory Capability](https://developer.apple.com/documentation/bundleresources/entitlements/com.apple.developer.kernel.increased-memory-limit) entitlement to ensure your app can allocate the required memory.
</Info>