Bootspring
HomeGet Started

ChatGPT Setup Guide

Using Bootspring-generated context with ChatGPT for AI-assisted development.

Overview#

While ChatGPT doesn't have native Bootspring integration, you can leverage the context files Bootspring generates to get better, more project-aware responses.

How It Works#

  1. Bootspring generates context - Creates CLAUDE.md with project details
  2. Share context with ChatGPT - Paste or upload the context
  3. Get project-aware responses - ChatGPT understands your codebase

Prerequisites#

  • ChatGPT account (Free or Plus)
  • Bootspring CLI installed locally
  • A project to work on

Setup#

Step 1: Install Bootspring#

npm install -g @girardmedia/bootspring

Step 2: Generate Context#

cd your-project bootspring init bootspring generate

Step 3: Prepare Context for ChatGPT#

1# View your context 2cat CLAUDE.md 3 4# Copy to clipboard (macOS) 5cat CLAUDE.md | pbcopy 6 7# Copy to clipboard (Linux) 8cat CLAUDE.md | xclip -selection clipboard 9 10# Copy to clipboard (Windows) 11type CLAUDE.md | clip

Using with ChatGPT#

Basic Usage#

  1. Start a new ChatGPT conversation
  2. Paste your project context
  3. Ask your questions

Example prompt:

I'm working on a software project. Here's the project context: --- [Paste CLAUDE.md contents here] --- Help me add a new API endpoint for user preferences.

ChatGPT Plus: Custom GPT#

Create a Custom GPT with your project context:

  1. Go to ChatGPT → Explore → Create a GPT
  2. Name it "My Project Assistant"
  3. In Instructions, add:
    You are a coding assistant for this specific project. Always reference the project context provided in Knowledge. Follow the coding patterns and conventions established in the project.
  4. Upload CLAUDE.md and bootspring.config.js to Knowledge
  5. Save and use for ongoing development

ChatGPT Plus: File Uploads#

Upload files directly:

  1. Click the attachment icon
  2. Upload CLAUDE.md
  3. ChatGPT will analyze and remember it

Effective Prompts#

Feature Development#

Project context: [paste CLAUDE.md] I need to implement user notifications. Based on our project structure and patterns: 1. Where should the notification logic go? 2. What database schema changes are needed? 3. Provide implementation code following our conventions.

Code Review#

Project context: [paste CLAUDE.md] Review this code for our project: ```typescript // paste code here

Check for:

  • Consistency with our patterns
  • Security issues
  • Best practices
### Debugging

Project context: [paste CLAUDE.md]

I'm getting this error: [paste error]

From this file: [paste relevant code]

What's causing it and how do I fix it?

### Architecture Decisions

Project context: [paste CLAUDE.md] Analysis report: [paste CODEBASE_ANALYSIS.md]

Should we switch from REST to GraphQL? Consider:

  • Our current architecture
  • Team expertise
  • Scaling requirements
## Best Practices ### Context Size Management ChatGPT has context limits. For large projects: ```javascript // bootspring.config.js module.exports = { context: { // Generate focused context include: ['src/**/*.ts'], exclude: ['**/*.test.ts', 'node_modules/**'], maxFileSize: 20000, maxContextSize: 50000, // ~12K tokens }, };

Selective Context#

For specific tasks, include only relevant parts:

# Get just architecture context bootspring context read architecture # Get just API context bootspring context read api

Refreshing Context#

Update ChatGPT when your project changes:

# Regenerate bootspring generate # Copy new context cat CLAUDE.md | pbcopy

Then update your Custom GPT or start a new conversation.

Comparison: Bootspring Agents vs ChatGPT#

Using Bootspring Agents#

# Direct expert consultation bootspring agent invoke security-expert "Review my auth implementation"

Pros:

  • Specialized prompts
  • Structured output
  • Integrated with project

Using ChatGPT with Context#

Acting as a security expert, review my authentication implementation. Project context: [CLAUDE.md] Code: [paste code]

Pros:

  • Conversational
  • Follow-up questions
  • Image/diagram support (Plus)

Workflow Examples#

Daily Development#

Morning:

bootspring health bootspring todo list

Development with ChatGPT:

My project context: [CLAUDE.md] Today's tasks: [from bootspring todo list] Help me implement task #1: "Add user preferences API"

Validation:

bootspring quality pre-commit

Learning Your Stack#

Project context: [CLAUDE.md] I'm new to this codebase. Explain: 1. The overall architecture 2. Key patterns used 3. How data flows through the app

Debugging Complex Issues#

# Generate analysis bootspring analyze
Project context: [CLAUDE.md] Analysis: [CODEBASE_ANALYSIS.md] I'm having performance issues with the user dashboard. The analysis shows [specific findings]. What should I investigate?

Limitations#

No Real-Time Access#

ChatGPT cannot:

  • Access your filesystem
  • Run commands
  • Modify files directly
  • See real-time code changes

Context Memory#

  • Free: Limited context per conversation
  • Plus: Better context retention
  • Custom GPTs: Persistent knowledge

No Code Execution#

ChatGPT generates code but can't:

  • Run tests
  • Execute builds
  • Validate output

Always validate locally:

bootspring quality pre-commit

Tips for Better Results#

1. Be Specific#

# Less effective "How do I add authentication?" # More effective "Following the patterns in my project context, implement JWT authentication with: - Access tokens (15 min expiry) - Refresh tokens (7 day expiry) - httpOnly cookies - Token rotation on refresh"

2. Reference Context#

"According to our tech stack (Next.js 14, TypeScript, Prisma), what's the best way to implement..."

3. Iterate#

# First prompt "Create the database schema for notifications" # Follow-up "Now create the API routes following our pattern in app/api/" # Follow-up "Add the frontend hooks using our React Query pattern"

4. Validate#

# Ask ChatGPT "Review this code for issues" # Then validate locally bootspring quality pre-commit bootspring agent invoke security-expert "Review recent changes"

Advanced Usage#

Code Interpreter (Plus)#

Upload multiple files for analysis:

  • CLAUDE.md
  • package.json
  • Key source files

ChatGPT can analyze relationships and suggest improvements.

DALL-E for Architecture#

Based on this project context: [CLAUDE.md] Create a diagram showing: - System architecture - Data flow - Service relationships

Plugins (Plus)#

Combine with plugins:

  • WebPilot: Research documentation
  • Code Interpreter: Analyze dependencies
  • Diagrams: Create architecture visuals

Next Steps#