Skip to main content
 /\_/\
( o.o )
 > ^ <

​
What is Splat?

Splat is an AI-powered debugging CLI tool that combines your compile and runtime errors with deep contextual analysis from your codebase. When an error occurs, Splat grabs context from every crevice of your project to deliver highly educated debug responses. Instead of manually tracing through stack traces and hunting down related files, Splat automatically:
  • Captures error output and stack traces
  • Parses affected files from the error
  • Optionally builds a dependency graph to find all related code
  • Sends everything to an AI model for instant analysis
  • Provides actionable fix recommendations
Created at CalHacks 11.0, Splat is perfect for whenever you have an error that needs to be quickly fixed, understood, and prevented in the future.
Development Status: Splat is an active hackathon project. The CLI command interface (splat squash) is currently under development. The core error analysis pipeline (in relational.py, process.py) is functional, but CLI integration is in progress. See the Contributing Guide to help complete the implementation.

​
Key Features

AI-Optimized Performance

Uses Groq’s extremely fast inference model to deliver instant debug responses with high accuracy

Zero Configuration

Works out-of-the-box globally in any project without configuration files or setup

Git Aware

Automatically respects .gitignore files to exclude sensitive information from analysis

Highly Contextual

Use the -r flag to grab all files related to your error stack through Nth-degree dependency analysis

​
How It Works

Splat intercepts your command execution and captures any errors that occur:
  1. Run your code - Execute your application through Splat instead of directly
  2. Error capture - Splat captures stdout, stderr, and the full error traceback
  3. Context gathering - Parses the stack trace to identify all involved files
  4. Dependency analysis (optional with -r flag) - Builds an adjacency graph to find related files
  5. AI analysis - Sends the error and context to Groq for intelligent debugging
  6. Get solutions - Receive structured explanations and actionable fix recommendations

​
Use Cases

Quickly fix missing parentheses, brackets, or other syntax issues with AI-guided solutions that explain why the error occurred.
Splat’s relational mode (-r) traces through your import statements to identify circular dependencies, missing modules, or path issues.
Debug FastAPI, Django, or Flask applications by capturing runtime errors with full request context.
Get clear explanations of type mismatches and validation errors with suggestions for proper type annotations.

​
Quick Example

Instead of running your Python script directly: 
python3 app.py
Run it through Splat: 
splat squash "python3 app.py"
When an error occurs, Splat automatically captures it, analyzes the relevant code, and provides AI-powered debugging assistance.

​
Get Started

Ready to supercharge your debugging workflow?

Installation

Install Splat using pip and set up your environment in minutes

Quick Start Guide

Go from installation to your first successful debug in under 5 minutes
Splat currently works best with Python projects. Support for additional languages is in development.

Build docs developers (and LLMs) love

Get started for freeTalk to us