SoloRoleplayer//

Download

Three things. One simple local lane.

For the Free version, you need SoloRoleplayer//, Cute LM, and the tested Qwen model. If you want the stronger lane, add our LoRA after that.

1 · The shell

SoloRoleplayer// Free

The offline game shell. This is the front end you play in.

Pick the platform that matches your machine. The Free lane is the offline Cute LM version.

1B · The API lane

SoloRoleplayer// BYOAI

For people who want to bring their own API instead of running the local Cute LM lane. Use this if you want ChatGPT, Claude, Gemini, or another supported API model.

This lane is for people bringing their own API. Best results if you turn up your max tokens to match the RP strength you picked at setup: Level 1, 2, or 3 posing style.

2 · The runtime

Cute LM

Our free local runtime. It runs your model privately on your own machine and is preset for SoloRoleplayer//.

Get Cute LM

Cute LM defaults to 127.0.0.1:1234 so SoloRoleplayer// can find it without weird setup.

3 · The tested model

Qwen 3.5 9B MLX

This is the one tested local model lane for the Free build.

mlx-community/Qwen3.5-9B-MLX-4bit
Open on Hugging Face

What the Free version is built for

One tested local model lane.

Model familyQwen3.5-9B-MLX-4bit
RuntimeCute LM
Port127.0.0.1:1234
Free version limitBuilt for one main LM lane so setup is easier and support is cleaner.
Private laneModel, runtime, saves, Shards, and local memory stay on your machine.
Internet useYou need internet to download the app, Cute LM, the model, and optional LoRA. After that, local play can stay offline.

Which version?

Pick the lane that fits how you want to play.

SoloRoleplayer// FreeBest for people who want the simple local lane: Cute LM + the tested Qwen 9B model + optional LoRA. More private. Less setup drift.
SoloRoleplayer// BYOAIBest for people who want to bring their own API. Use this if you already have access to ChatGPT, Claude, Gemini, or another supported API model.

Optional upgrade

Our LoRA for better roleplay discipline.

We offer a WordBlock Labs LoRA trained for SoloRoleplayer//. It almost completely stops head-hopping and persona loss in the tested Qwen 9B lane.

Need help?

Use the simple setup guides.

If you want the easy path, use these in order:

  1. How to download the tested Qwen model
  2. How to start Cute LM
  3. How to add the SoloRoleplayer// LoRA

Support the solo dev.

WordBlock Labs is a solo dev shop. If you want more weird, useful, private text tools like this, support helps keep them alive.

Support on Ko-fi