telescopeai/autodebug

AI-powered auto-debug & auto-fix system for Laravel. Monitors Telescope exceptions, analyzes them with AI, generates fixes, and creates GitHub PRs automatically.

dev-main 2026-03-12 19:34 UTC

Requires

• php: ^8.1
• illuminate/console: ^10.0|^11.0|^12.0
• illuminate/database: ^10.0|^11.0|^12.0
• illuminate/http: ^10.0|^11.0|^12.0
• illuminate/routing: ^10.0|^11.0|^12.0
• illuminate/support: ^10.0|^11.0|^12.0

Requires (Dev)

• orchestra/testbench: ^8.0|^9.0|^10.0
• phpunit/phpunit: ^10.0|^11.0

MIT

• TelescopeAI

• debugging
• github
• laravel
• automation
• ai
• Gemini
• telescope
• openai
• claude
• ollama
• auto-fix

README

AI-powered auto-debug & auto-fix for Laravel. Monitors Telescope exceptions, analyzes them with multiple AI providers (OpenAI, Claude, Gemini, or local Ollama), generates code fixes, and creates GitHub PRs automatically.

✨ Features

• 🔍 Automatic Exception Detection — Polls Telescope for new exceptions
• 🧠 Multi-AI Engine — Supports OpenAI, Anthropic (Claude), Google Gemini, and Ollama (Local/Free)
• 🛠️ Auto-Fix Generation — AI suggests code patches with search/replace
• 🚀 GitHub PR Creation — Pushes fixes as PRs with detailed descriptions
• �️ Terminal Diff Preview — View suggested changes directly in your console
• �📊 Web Dashboard — View exceptions, AI analysis, confidence scores, and PR status
• 🔔 Notifications — Slack, email, and database logging
• 🔒 Safety Guards — Protected paths, deduplication, rate limiting, confidence thresholds

📦 Installation

1. Require the package

composer require telescopeai/autodebug

2. Run the install command

php artisan autodebug:install

3. Configure your .env

Option A: Local Ollama (Free & Private)

Best for internal development. No API keys required.

AUTODEBUG_AI_PROVIDER=ollama
AUTODEBUG_OLLAMA_BASE_URL=http://localhost:11434
AUTODEBUG_OLLAMA_MODEL=deepseek-coder:6.7b

Option B: Google Gemini

Highly capable with generous free tiers.

AUTODEBUG_AI_PROVIDER=google
AUTODEBUG_GOOGLE_API_KEY=your-gemini-key
AUTODEBUG_GOOGLE_MODEL=gemini-2.0-flash

Option C: OpenAI / Anthropic (Claude)

Professional grade models.

# For OpenAI
AUTODEBUG_AI_PROVIDER=openai
AUTODEBUG_OPENAI_API_KEY=sk-your-key-here

# For Anthropic
AUTODEBUG_AI_PROVIDER=anthropic
AUTODEBUG_ANTHROPIC_API_KEY=sk-ant-your-key-here

4. GitHub Configuration

Required only if you want automatic PR creation.

AUTODEBUG_GITHUB_ENABLED=true
AUTODEBUG_GITHUB_TOKEN=ghp_your-github-token
AUTODEBUG_GITHUB_OWNER=your-org-or-username
AUTODEBUG_GITHUB_REPO=your-repo-name

🚀 Usage

CLI Commands

# Run analysis (Dry run doesn't create PRs)
php artisan autodebug:analyze --dry-run

# 🔥 See the file changes in terminal
php artisan autodebug:analyze --dry-run --diff

# Force analysis even if recently analyzed
php artisan autodebug:analyze --force

⚙️ Configuration

Option	Default	Description
ai.provider	openai	openai, anthropic, google, or ollama
analysis.min_confidence_for_pr	75	Minimum AI confidence to create a PR
analysis.max_calls_per_hour	10	Rate limit for AI API calls
analysis.dry_run	false	Global dry run mode
safety.protected_paths	[...]	Files the AI is never allowed to touch

🧪 Prerequisites

• PHP 8.1+
• Laravel 10, 11, or 12
• Laravel Telescope
• GitHub Personal Access Token (for PRs)

📄 License

MIT License. See LICENSE for details.