▄▄▄ ▄▄▄ ▄▄▄▄▄▄▄
████▄ ▄████ ███▀▀███▄
███▀████▀███ ▄█▀█▄ ███▄███▄ ▄███▄ ███▄▄███▀ ▀▀█▄ ▄█▀▀▀ ▄█▀█▄
███ ▀▀ ███ ██▄█▀ ██ ██ ██ ██ ██ ███ ███▄ ▄█▀██ ▀███▄ ██▄█▀
███ ███ ▀█▄▄▄ ██ ██ ██ ▀███▀ ████████▀ ▀█▄██ ▄▄▄█▀ ▀█▄▄▄
MemoBase — Memory for your codebase
MemoBase is a production-grade, offline-first codebase memory system that provides intelligent code analysis, search, and visualization capabilities. Built with performance and modularity in mind.
- Streaming Processing: No full repository loading
- Parallel Parsing: Multi-core file processing
- Intelligent Indexing: Fast semantic search
- Graph Visualization: Code relationship mapping
- CLI & TUI: Flexible interface options
- Incremental Updates: Efficient change detection
- Offline-First: No external dependencies
pip install memobase# Initialize in your project
memobase init
# Build the memory index
memobase build
# Search your codebase
memobase ask "How does authentication work?"
# Launch TUI
memobase tuiMemoBase follows a strict modular architecture:
BUILD PIPELINE:
Filesystem Scanner → Parser → Memory Builder → Storage → Index → Graph
QUERY PIPELINE:
Input → Intent Classifier → Retriever (Index + Graph) → Formatter → Output
TUI LOOP:
Input → State Manager → Renderer (partial updates only)
- parser/: Tree-sitter based code parsing
- memory/: Memory unit construction
- storage/: Persistent data storage
- index/: Search indexing
- graph/: Relationship graph building
- query/: Query processing and retrieval
- analysis/: Code analysis tools
- incremental/: Change detection and updates
- cli/: Command-line interface
- tui/: Terminal user interface
- infrastructure/: Core utilities
- tests/: Comprehensive test suite
- docs/: Documentation
- <2s query latency on large codebases
- Streaming processing with constant memory usage
- Parallel parsing with ProcessPoolExecutor
- Lazy loading throughout the system
See the docs/ directory for detailed documentation:
MIT License - see LICENSE file for details.