---
title: Project Task Planner
emoji: 📋
colorFrom: red
colorTo: indigo
sdk: docker
app_port: 7860
pinned: true
license: mit
short_description: Project plans with a small local LLM on CPU, no API
models:
  - Qwen/Qwen2.5-0.5B-Instruct
---

# Project Task Planner

Turns a project idea into a structured plan: tasks, subtasks, owners, hour estimates and a rough duration for your team size.

**How it works**

1. Pick one of six project types (e-commerce, mobile, SaaS, AI/ML, enterprise, blockchain). The plan comes from tested templates, so it is instant and always well-formed.
2. Optionally click **Ask the local model**. [Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct) runs on this Space's free CPU and adds risks and missing tasks specific to your project. Nothing is sent to an external API.
3. Tick off subtasks and export the plan as Markdown, CSV or JSON.

**Why a 0.5B model?** It is small enough to run on a 2-vCPU machine without a GPU. That is the same constraint as running AI locally for a small business that does not want its data leaving its own hardware.

**Run it yourself**

```
docker build -t planner . && docker run -p 7860:7860 planner
```

Author: [Chad Saglam](https://chadev.ch) · MIT license