Skip to content
  • cli

BuddhiAI CLI overview

BuddhiAI CLI turns a software repository into a local code graph and a set of tools and workflows for coding agents. The package requires Python 3.10 or newer and installs the buddhi command. The optional mcp extra also installs the buddhi-mcp stdio server.

Main parts

Section titled “Main parts”

Code graph

Section titled “Code graph”

buddhi init and buddhi generate parse supported source with tree-sitter and write graph artifacts under .buddhi/graphs/. The graph records files, directories, classes, functions, methods, imports, calls, and containment relationships. tree-graph.db backs graph queries; tree-graph.json and tree-graph.html provide visualization artifacts.

The graph supports buddhi graph commands for statistics, community analysis, dependency paths, import cycles, and blast-radius investigation. See Graph analysis.

MCP server

Section titled “MCP server”

buddhi-mcp runs over stdio and provides tools for code search, compressed file reading, project memory, contrast checks, and accessibility verification. See MCP server and tools.

OKF documentation and memory

Section titled “OKF documentation and memory”

buddhi docs generates documentation in Google Open Knowledge Format (OKF) from the graph and can build or serve the generated Pelican site. buddhi memory manages source-aware concepts and memory indexes. These operate on the target repository; see Documentation generation and Project memory.

Antigravity harness

Section titled “Antigravity harness”

buddhi init can scaffold .agents/ with workflows, specialist agents, rules, hooks, templates, and a memory index. The included workflows support specification, planning, task breakdown, implementation, and verification. See Agent harness.

Supported languages

Section titled “Supported languages”

The CLI parses Python, JavaScript, TypeScript and TSX, Go, Rust, C#, Java, Kotlin, and Swift. Import and call resolution is best-effort; see the accuracy notes in the CLI README.

Common setup

Section titled “Common setup”
  1. Install BuddhiAI CLI with the MCP extra using the installation guide.
  2. From the repository root, run buddhi init.
  3. Open the repository in Google Antigravity to use the generated harness and workspace MCP configuration.
  4. Run buddhi generate after source changes to refresh the graph.

Browse the command reference for command syntax and the MCP guide for connecting other compatible clients.