DOCSEARCHbyAlgolia
Submit Docs

MemoV Documentation

website·Indexed 32 days ago

https://docs.memov.ai

A Git-powered version control layer designed for AI-assisted development. MemoV captures AI coding history using a shadow .mem timeline and provides a CLI, a Web UI for exploration, and MCP tools for server-side operations.

Tokens
5.5K
Snippets
49
Records
67
Agent score
57%

What's inside MemoV

  1. Overview of MemoV features and capabilities

    MemoV (VibeGit) provides traceable, Git-powered memory for AI coding. It automatically versions prompts, agent plans, and code diffs in a shadow .mem timeline, ensuring the primary .git history remains clean. Key capabilities include:

    • Branch Exploration: Branch, rollback, and replay interactions without erasing history.
    • Cross-session Persistence: Memory persists across different AI sessions.
    • Local-first Privacy: Operates locally with no database; supports .memignore to exclude specific files.
    • Visual UI: A web-based interface for browsing history and diffs available at http://localhost:38888.

  2. Overview of MemoV memory layer

    MemoV is a memory layer designed for AI coding agents. It provides traceable, Git-powered memory that automatically versions prompts, context, and code diffs. Key features include persistent memory across sessions, branch exploration for different solutions, full traceability from prompt to code change, a shadow .mem timeline to prevent .git pollution, and a visual history of the AI coding journey.

  3. Overview of MemoV Architecture

    MemoV is a Git-powered version control layer for AI-assisted development. It captures user prompts, agent plans, and code changes as atomic units, storing them in a shadow .mem/memov.git repository. This system is designed to be completely independent of the project's standard .git/ directory, allowing conversational context to be tracked at per-interaction granularity without polluting the production history.

  4. Initialize MemoV storage

    Run the mem init command to set up the storage architecture. This operation creates the .mem/ directory, initializes a bare Git repository at .mem/memov.git/, generates a default .memignore file, and creates an empty branches.json file.
    mem init

  5. Understand the MemoV Shadow Timeline architecture

    MemoV uses a parallel version control system located in .mem/memov.git. This is a bare Git repository that operates independently from the project's standard .git directory. This design prevents pollution of the production Git history while allowing MemoV to capture complete AI interaction context using Git notes to store enhanced metadata (conversational context) alongside standard Git content-addressing.

  6. Explore AI coding history using Web UI features

    The MemoV Web UI includes several views for analyzing AI-generated code changes:

    • Timeline View: Chronological list of interactions showing commit hashes, timestamps, prompts, and branch labels.
    • Tree View: Hierarchical folder structure with file status indicators (modified, clean, untracked) and path filtering.
    • Branch Filtering: Ability to view all branches (e.g., main, develop/0), filter the timeline by branch, and visualize branch relationships.
    • Diff Viewer: Side-by-side comparison with syntax highlighting, showing additions (green) and deletions (red).

  7. Install the mem CLI from source

    To install the mem CLI from the source code, clone the repository and use uv for dependency synchronization and installation.
    git clone https://github.com/memovai/memov.git
cd memov
uv sync
uv pip install -e .
mem --help

  8. Install mem-mcp with RAG support

    RAG tools require the [rag] extras installation. Depending on your environment, use the following commands:
    # For Claude Code
claude mcp add mem-mcp --scope project -- uvx --from "git+https://github.com/memovai/memov.git[rag]" mem-mcp-launcher stdio $(pwd)

    For VS Code / Cursor

    Update the --from argument to: "git+https://github.com/memovai/memov.git[rag]"

  9. Integrate MemoV MCP with Claude Code

    To integrate MemoV with Claude Code, run the following command in your project root directory. This provides automatic updates on agent restart.
    claude mcp add mem-mcp --scope project -- uvx --from git+https://github.com/memovai/memov.git mem-mcp-launcher stdio $(pwd)

  10. Install the mem CLI using package managers

    The mem CLI is available via Homebrew for macOS/Linux and APT for Debian/Ubuntu systems.
    # Homebrew (macOS/Linux)
brew tap memovai/mem
brew install memov

# APT (Debian/Ubuntu)
echo "deb [trusted=yes] https://memovai.github.io/memov/apt stable main" | sudo tee /etc/apt/sources.list.d/mem.list
sudo apt update
sudo apt install mem

  11. Capture AI interaction context using snap

    The snap tool integrates AI agents with MemoV's versioning system by capturing four key parameters for every interaction: the original user prompt, the AI's complete response, the agent's high-level change plan (list of strings), and a comma-separated list of modified file paths. MemoV handles these interactions in three tiers: recording metadata for prompt-only interactions, detecting manual user edits by comparing actual changes against the declared files_changed list, and classifying files as untracked, modified, or clean.

  12. Access the MemoV Web UI

    The MemoV Web UI provides a timeline view, branch filtering, and a diff viewer. You can open it by typing "use mem ui" in the AI chat. The UI is hosted locally at http://localhost:38888.