Sukhia

SukhiaV1.0

SukhiaV1.0 is a Qwen3-8B-based coding assistant configuration for local Ollama use. It is customized with a system prompt and technology guidance; it is not a weight-fine-tuned model and is not Qwen3-Coder.

Supported technologies

PHP 8.3+, Laravel, Node.js, NestJS, React, Next.js, Python, Docker, Terraform, and TypeScript.

Install with Ollama

Requirements: Ollama and enough local memory for qwen3:8b.

git clone https://github.com/dipeshsukhia/Sukhia.git
cd Sukhia
bash scripts/create-ollama.sh

Or run the steps manually:

ollama pull qwen3:8b
ollama create sukhia -f Modelfile
ollama run --think=false sukhia

For the complete Ollama and Open WebUI setup, see docs/LOCAL-SETUP.md.

To start the model and WebUI together:

bash scripts/start-sukhia.sh

To stop the services while preserving downloaded models and WebUI data:

bash scripts/stop-sukhia.sh

Validate and package

python3 scripts/validate.py
bash scripts/package-release.sh v1.0.0

Scope and limitations

SukhiaV1.0 provides coding guidance and generated examples. Review and test all output. It does not execute code, access private repositories, guarantee framework-version accuracy, or replace security review. No private or proprietary training data is included.

Upstream attribution

The base model is Qwen/Qwen3-8B. Follow the upstream license and notices when redistributing this configuration.