Quick Start

NestWeaver parses 32 languages via Tree-sitter — JS, TS, Go, Python, Rust, Java, C/C++, Lua, Scala, Elixir, Zig, Vue, Svelte, and more — and exposes 40 MCP tools for AI agents. Here’s how to go from install to your first query.

Set up your AI tools

NestWeaver integrates with 16 AI tools out of the box. Run setup to auto-configure them:

nestweaver setup
# NestWeaver Setup
# ────────────────────────────────────────
# ✓ Claude Code — .claude/settings.json — MCP server configured
# ✓ Cursor — .cursor/mcp.json — MCP server (lite: 6 tools)
# ...

This writes the correct MCP server configuration for each detected tool (Claude Code, Cursor, Codex, Gemini CLI, Copilot CLI, Aider, Kiro, and more). Run nestweaver setup --force to regenerate files even if you’ve customized them.

Index your codebase

Point NestWeaver at a repository to build the knowledge graph:

nestweaver index --repo .
# Indexing /path/to/repo → ./nestweaver.lbug

NestWeaver auto-detects the repo root from .git, parses every supported file, resolves cross-file references with confidence scoring, and writes everything to a local database.

To include git history analysis (co-change mining and recency scoring):

nestweaver index --repo . --with-git-activity

Start querying

Ask structural questions about your code:

nestweaver context processPayment

This uses Personalized PageRank to surface the symbols most relevant to your query — call sites, dependencies, type definitions, and more — without reading raw source files.

Search across symbols and notes:

nestweaver search "main"
# Found N symbol(s) matching 'main':

Watch for changes

Keep the index up to date as you code:

nestweaver watch

This watches the filesystem and re-indexes on changes with debouncing, so your AI tools always have a current graph.

Next steps

• Your First Query — a deeper walkthrough of querying the graph
• MCP Tools — the full list of 40 tools available to AI agents
• Configuration — customize indexing, token budgets, and more