ChatGPT Prompts for Coding Projects: 12 to Plan and Finish

ChatGPT prompts for coding projects help most when they size the project to your real skills and hours, then break it into milestones you can test yourself. The 12 prompts below cover picking an idea, writing a spec, planning the build, testing, getting unstuck, and presenting the finished work. Paste your stack and time budget every time, or you will get a plan built for someone else.

Why most side projects stall

Projects die when the first version is too big, the tools are new all at once, or there is no clear finish line. A good prompt fixes all three: it asks for a first version that fits half your time, keeps to tools you named, and gives every milestone a demo and a done check. Tell the model what you have already built, because "beginner" means very different things.

Pick and scope the idea

1. Three ideas from my skills

I know [languages and what I built], I have [hours] per week for [weeks], and my goal is [goal]. Suggest three projects sized to that time, the concept each teaches, and which one you recommend.

2. Solve my own problem

Here is an annoying problem from my life: [problem]. Turn it into a small app idea I can build with [stack], with what is in scope for version one and what waits.

3. Scope cutter

Here is my project idea, [paste]. Cut it down to a version I can finish in [hours] total, and list what moves to later.

Spec, structure, and data

4. User stories and spec

Write 5 to 8 user stories, an out-of-scope list, and the screens or commands for this project: [idea]. Keep it to version one.

5. Folder layout

Give me a file and folder layout for [project] in [stack], with one line per file explaining its job.

6. Data model

Design the tables or data structures for [project] with field names, types, and relationships, and explain which constraint stops bad data.

Build week by week

7. Milestones with done checks

Split [project] into weekly milestones for [weeks] weeks. Each needs a demo I can show and a done check I can test myself.

8. First 90 minutes

Plan my first 90-minute session for [project] in [stack], ending with something that runs. Give the commands to set up the environment.

9. Test plan

List the unit tests and one end-to-end check for [feature], and show how to run them from the command line in [stack].

Get unstuck and ship

10. Error explainer

Here is my error and the code around it: [paste]. Explain what the error means in plain words, the most likely cause, and how to confirm it before changing code.

11. Code review

Review this file, [paste], like a senior engineer reviewing a junior's pull request: bugs, unclear names, missing tests, and one thing done well.

12. README that sells it

Write a README outline for [project] that shows what it does, the tech used, how to run it and its tests, and what I would build next.

Worked example

A learner with about six months of Python, a CLI to-do app, and a weather script wanted a first backend portfolio project in Flask and SQLite, with 6 hours a week for 4 weeks. Their real problem: climbing friends lost track of route sends in a group chat. The flagship prompt suggested three ideas and recommended a small Send Log web app that friends can open on their phones.

The plan included user stories, a three-table schema, weekly milestones with done checks, the first 90 minutes ending with a running page, tests to run with pytest, and fallbacks if login or deployment stalled. Get the full coding project build plan prompt.

Related coding prompts

When you start writing code, the code generation prompt with spec and tests keeps output tied to your plan. New to the language? Try the ChatGPT prompt for learning Python. For errors, use the ChatGPT prompt for debugging, and for front ends see ChatGPT prompts for web development.

Find more in the Coding prompts hub, browse all prompts, or share your own prompt.