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

# Mako-8B Operator

> The compact model optimized for streaming and low-latency interaction.

## Overview

| Property           | Value                    |
| ------------------ | ------------------------ |
| **Parameters**     | 8 billion                |
| **Architecture**   | Decoder-only transformer |
| **Context Window** | 4,096 tokens             |
| **Streaming**      | Supported                |
| **Request format** | `model: "operator"`      |

The Operator is Mako's compact model, optimized for streaming responses and real-time interaction. It shares the same personality, tool suite, and agentic capabilities as the 32B Conductor, in a smaller package.

<Note>
  The 32B Conductor is now the default and recommended model for all use cases. The Operator is available for self-hosted deployments where lower resource usage is preferred.
</Note>

## Streaming events

When streaming, the Operator emits additional SSE events beyond standard content deltas:

* `tool_start` — emitted when a tool begins execution
* `tool_trace` — emitted when a tool completes, with the result
* `agent_text` — intermediate model text during tool rounds

See [Streaming](/api-reference/streaming) for details on consuming these events.

## Self-hosting

The Operator model weights are available on [Hugging Face](https://huggingface.co/deepmako/Mako-8B-Operator) in GGUF format. Run it locally with Ollama or llama.cpp for development and testing.

See [Self-Host](/self-host/gateway) for setup instructions.