The Problem Batho Solves
AI coding agents are powerful — but they burn tokens reading files and hallucinate when context is thin. Batho gives your agent a structured code graph so it works smarter, not harder.
Slash Token Costs
Your agent queries a graph instead of reading entire files. 10x fewer tokens per task — no more dumping your repo into the LLM.
Eliminate Hallucinations
Deterministic, tree-sitter-parsed relationships. Your agent gets facts, not guesses — zero hallucinations on structural queries.
Agent Superpowers
Bug tracking, security audits, refactoring, code review — your agent handles more, accurately. When cost and quality are solved, automation widens with imagination.
How It Works
From source code to AI-ready graph in four steps
Build
Parse your entire codebase into a structured hypergraph with cross-file symbol resolution.
batho build --root .Patch
Apply incremental updates 10-100x faster than full re-indexing using native content hashing.
batho patch --root .Export
Export a transportable ZIP artifact by default, or use --json for LLM-optimized views.
batho export --root .Verify
Execute integrity checks and diagnostic routines to repair database corruption or inconsistencies.
batho fix --dry-runEverything You Need
From multi-language parsing to zero-copy storage, Batho gives your agent the tools to understand any codebase.
Works With Your Stack
Python, TypeScript, Rust, Go, Java, and more. Batho parses 40+ languages so your AI agent understands your entire codebase — no blind spots.
Learn more →Track Codebase Evolution
Versioned snapshots and incremental diffing let your agent understand what changed and why — across every commit.
Learn more →Zero-Copy Performance
Apache Arrow IPC storage means zero-copy, memory-mapped reads. Your agent queries the graph instantly.
Learn more →Reconstruct Anything
Reconstruct any file byte-for-byte from the graph. Cryptographic hash integrity means nothing is lost.
Learn more →Zero Config to Start
Batho runs with zero config. Customize with a single YAML file when you need control.
Learn more →MCP-Native
Works with Claude Code, Cursor, Windsurf, Antigravity, Gemini CLI, Cline, OpenCode, and Aider. One MCP config, zero hassle.
Learn more →What Your Agent Can Do
When cost and quality are solved, automation widens with imagination. Here's what becomes possible.
Bug Tracking
Trace root causes across files. Your agent follows the graph — not grep — to find where bugs originate.
Security Audits
Map attack surfaces, trace data flows, and identify vulnerable patterns — structurally, not heuristically.
Refactoring
See every dependency before you rename. Your agent knows the blast radius of every change.
Code Review
Review PRs with full context. Your agent understands what changed, what depends on it, and why it matters.
Onboarding
New team members ask your agent questions. It answers with structural facts — not file dumps.
CI/CD Integration
Build the graph in CI, export an artifact, and ship it with your deploy. Agents query locally — no server needed.
Works With Your AI Agent
Batho connects to any MCP-compatible coding agent. One config, zero hassle. Your agent gets a code graph instead of a file dump.
Claude Code
Anthropic's terminal-first coding agent with the deepest MCP integration.
- 1M context window + Batho graph = near-zero hallucination on structural queries
- Hook system can trigger batho patch automatically after edits
- 10 MCP tools available as mcp__batho__* in every session
claude mcp add batho -- batho mcpGet Started in Seconds
Two paths to graph-powered coding. Pick the one that fits your workflow.
Setup with your AI Agent
1 command · ~30 secondsDownload the skill file and give it to your agent. It installs Batho, builds the graph, and configures MCP for all detected clients.
Full guidecurl -O https://raw.githubusercontent.com/sageoz/batho/main/SKILL.mdThen tell your agent: Read SKILL.md and set up Batho for this repo
Manual CLI Setup
3 commands · ~1 minutePrefer the terminal? Three commands — install, build, start MCP server. No agent required.
CLI quick startpip install batho && batho build --root . && batho mcpNo agent needed — run everything from your terminal. batho mcp starts the MCP server.
Give your AI agent a code graph — not a file dump.
Index in one command. Connect via MCP. Your agent gets structural intelligence with zero-copy speed, 10x fewer tokens, and zero hallucinations.
- One command to build
- Works with 8+ AI agents
- No server required