By anortham
Code intelligence server: search, navigation, and refactoring across 34 languages
Use before making changes to existing files. Routes to Julie's edit_file, rewrite_symbol, and rename_symbol tools without a Read + Edit loop. Trigger on fix, change, update, modify, refactor, rename, replace, add, remove, move, or any task involving an existing file.
Orient on a new codebase area using get_context for token-budgeted exploration. Use when the user asks "what does this module do", "explain this part", or wants to understand an unfamiliar area before making changes.
Analyze what would break if a symbol changes: finds callers, groups by risk level, assesses impact. Use when the user asks about blast radius, who uses a symbol, or is planning a refactor.
Fetch web pages and index them locally for token-efficient research. Use when you need to read documentation, articles, or any web content. Fetches via browser39, saves as markdown, indexes via Julie's filewatcher, then uses Julie tools for selective reading. Requires browser39 (binary release from GitHub).
Admin access level
Server config contains admin-level keywords
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin that installs Julie, a Rust-based code intelligence MCP server. Julie gives AI coding agents LSP-quality search, navigation, and refactoring across 34 programming languages.
This package also ships manual install paths for the Codex CLI and OpenCode — see Codex CLI install and OpenCode install below.
For full documentation on Julie's tools, capabilities, and supported languages, see the Julie repository.
Install uv so Julie can auto-provision Python 3.12 and GPU-accelerated embeddings. You do not need to install Python yourself. If uv is missing or embeddings cannot start, Julie still provides keyword search and code navigation; embedding-backed features are disabled until the sidecar is available.
# macOS
brew install uv
# Windows (PowerShell)
winget install --id=astral-sh.uv -e
# Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
To enable the /web-research skill for fetching and indexing web content, install browser39:
# Download the latest release for your platform from:
# https://github.com/alejandroqh/browser39/releases
This is optional. All other Julie features work without browser39.
# Add the Julie plugin repository as a marketplace
/plugin marketplace add anortham/julie-plugin
# Install the plugin (user scope, available across all projects)
/plugin install julie@julie-plugin
You can also scope the installation to a specific project:
# Project scope (shared with team via version control)
/plugin install julie@julie-plugin --scope project
git clone https://github.com/anortham/julie-plugin.git
# Install as a plugin
claude plugin install /path/to/julie-plugin
# Or load from local directory (useful for development)
claude --plugin-dir /path/to/julie-plugin
fast_search, get_symbols, deep_dive, fast_refs, call_path, get_context, blast_radius, rename_symbol, manage_workspace, edit_file, rewrite_symbol)/editing, /explore-area, /impact-analysis, /web-research)The plugin intentionally does not install behavioral hooks. Julie's MCP tools and skills are the durable integration surface; hook-based coaching created too much client-specific configuration drift.
On first launch, Julie extracts a pre-built native binary, prepares the embedding sidecar when uv is available, detects your GPU, and indexes your codebase. Subsequent sessions load the cached index instantly with incremental updates for changed files.
The plugin ships an installer that wires Julie's skills into the Codex CLI. It does not touch any config.toml; it prints the exact MCP registration command for your checkout after install.
Prerequisites: Node.js 18+, Codex CLI installed.
Install:
# From a julie-plugin checkout (or anywhere — the script resolves paths from its own location)
node bin/install-codex.cjs
This is idempotent and adds:
skills/<name>/ to ~/.codex/skills/julie-<name>/ for all 4 skillsDuring install and uninstall it also removes legacy Julie hook entries and the old <!-- julie-precedence start/end --> AGENTS block if they exist.
Register the MCP server in Codex (one-time, you do this manually):
User-scope:
codex mcp add julie -- node /absolute/path/to/julie-plugin/hooks/run.cjs
Codex CLI and Codex Desktop do not send MCP roots, so Julie uses the process cwd/startup hint. If a desktop app starts Julie from the wrong directory, set JULIE_WORKSPACE in that app's MCP config or launch from the repo root:
[mcp_servers.julie]
command = "node"
args = ["/absolute/path/to/julie-plugin/hooks/run.cjs"]
env = { "JULIE_WORKSPACE" = "/absolute/path/to/your/project" }
Uninstall:
node bin/install-codex.cjs --uninstall
Removes the Julie skill symlinks and any legacy Julie hook/AGENTS artifacts; leaves your other Codex configuration intact.
The plugin ships a parallel installer that wires Julie's skills into OpenCode. It does not touch opencode.json; it prints a ready-to-copy MCP block after install.
Prerequisites: Node.js 18+, OpenCode installed.
Install:
node bin/install-opencode.cjs
This is idempotent and adds, all under ~/.config/opencode/:
npx claudepluginhub anortham/julie-plugin --plugin julieCross-client MCP memory with checkpoints, recall, briefs, and standups for AI-assisted development
Miller-powered development workflow with subagent-driven execution, inline review, and Miller-first codebase orientation.
Agent memory persistence - remember and recall across sessions
Composite workflows that orchestrate Julie (code intelligence), Sherpa (workflow guidance), and Goldfish (persistent memory) for powerful development patterns. Includes TDD Powerhouse, Bug Detective, Smart Session Start, and more.
Local SQLite-backed code intelligence, content search, and workspace freshness for coding agents.
Full AI context layer over MCP — tree-sitter code-map, document RAG (PDF/Office/HTML/email + OCR + reranker), shared agent memory, on-demand web crawl, git history + blame + per-symbol diff. 300+ languages, 8 coding-agent harnesses, content-addressed Fjall + LanceDB.
Semantic code intelligence -- symbol navigation, cross-reference analysis, and structural code understanding powered by tree-sitter
Precise local semantic code search via MCP. Indexes your codebase with Go AST parsing, embeds with Ollama or LM Studio, and exposes vector search to Claude through an MCP server — no cloud, no npm.
Tree-sitter backed code indexing and retrieval for Claude Code. Precise function extraction, cross-file callers, test discovery, and scope-aware grep.
Fast and token-friendly code reading for AI coding agents. Symbol-aware MCP tools that replace cat/grep with ~85% fewer tokens, sub-millisecond search, and a raw fallback that preserves cat/grep parity byte-for-byte.
Local-first hybrid semantic code search with dependency graphs and cross-repo search. Powered by CocoIndex, Tree-sitter, pgvector, and Ollama.