mufeedvh/code2prompt
Code2prompt
code2prompt is a codebase-to-prompt converter and LLM context generator that transforms source code and directory structures into formatted text blocks for large language models. It functions as both a utility for generating prompts and an AI agent context server that exposes codebase files and metadata to coding assistants via a standardized server protocol.
The tool distinguishes itself through git-aware capabilities, integrating commit messages and branch diffs to provide version control context for AI-generated code changes. It also utilizes the Model Context Protocol to allow external AI assistants to programmatically query project files.
The system manages context through glob-based file filtering and gitignore rule adherence to refine source file selection. It includes a token counter to ensure the aggregated codebase fits within a model's context window and employs custom templates to define the structural layout of the final output.
Features
- Repository-to-Prompt Converters - Transforms entire directory structures and source files into formatted text blocks ready for large language model ingestion.
- Repository-to-Prompt Converters - Transforms entire directory structures and source files into a single formatted text block for LLM context.
- AI Coding Assistant Integrations - Exposes the codebase through a standardized server protocol for programmatic analysis by AI coding assistants.
- AI Integration Protocols - Implements a standardized protocol to allow AI agents to programmatically access and query project files.
- Context Preparation Utilities - Processes and formats raw source code and directory structures into structured context for language models.
- LLM Context Preparation - Aggregates source code and directory structures into a single formatted text block for LLM prompts.
- Model Context Protocol Servers - Provides a server-side implementation of the Model Context Protocol to expose codebase files and metadata.
- Context Aggregators - Flattens multi-file directory structures into a single formatted text stream for efficient LLM ingestion.
- Model Context Protocol Integrations - Exposes codebase context through the Model Context Protocol for programmatic querying by external AI assistants.
- Token Context Limiting - Tracks token volume to ensure the processed codebase fits within the target model's maximum context window.
- Git Repository to Prompt Converters - Extracts git commit history and diffs to provide temporal version control context within AI prompts.
- Prompt Templates - Provides a templating system to define the structure and layout of the aggregated codebase prompt.
- Codebase Analysis Tools - Curates specific sets of source files using templates and globs for structured AI-driven documentation.
- Codebase Size and Token Count Reporters - Calculates the total token count of the generated text to ensure it fits within the model's context window.
- Linting File Filters - Uses glob patterns and gitignore rules to refine and exclude specific files from the AI context.
- Git Context Injectors - Includes commit messages and branch diffs in prompts to support automated pull request and commit generation.
- Glob Pattern Resolvers - Uses glob patterns and gitignore rules to filter and select relevant source files for processing.
- Prompt Templates - Employs customizable text templates to ensure consistent formatting of aggregated code and metadata in prompts.
- Structured File Filtering Systems - Applies structured .gitignore and glob rules to filter files during the codebase ingestion process.
- Code and Context Tools - CLI tool to convert codebases into single LLM prompts.
