tree-sitter-language-pack

Parse and extract code intelligence from 300+ programming languages with tree-sitter — structure, imports, symbols, and syntax-aware chunking — using the local ts-pack CLI in your agent.

Install

From the marketplace (recommended)

Pending review for official Claude marketplace.

Self-host:

/plugin marketplace add kreuzberg-dev/plugins
/plugin install tree-sitter-language-pack@kreuzberg

Binary requirement

The bundled MCP launcher (scripts/mcp-launch.sh) resolves a ts-pack binary automatically on first run: it reuses one already on PATH, then tries npx/uvx, then Homebrew, then a prebuilt download from the tool's latest GitHub release. No manual install is required to use the MCP server.

To install the ts-pack CLI yourself:

brew install kreuzberg-dev/tap/ts-pack
# or run it without a persistent install (the CLI proxy package self-installs the binary):
npx @kreuzberg/ts-pack-cli --help
uvx --from ts-pack-cli ts-pack --help
# or download a prebuilt binary from the latest GitHub release:
#   https://github.com/kreuzberg-dev/tree-sitter-language-pack/releases/latest
# or build from source (binary is installed as `ts-pack`):
cargo install --git https://github.com/kreuzberg-dev/tree-sitter-language-pack ts-pack-cli

Parser libraries download on demand the first time a language is used. For offline or CI runs, prefetch them:

ts-pack download --all          # every language
ts-pack download python rust    # specific languages

SDKs (optional)

pip install tree-sitter-language-pack                  # Python
npm install @kreuzberg/tree-sitter-language-pack       # Node.js / TypeScript

Rust, Ruby, Go, Java, C#, PHP, Elixir, and WebAssembly bindings are also published.

Skills shipped

| Skill | Trigger | |-------|---------| | tree-sitter-language-pack | Parse and extract code intelligence from 306 languages. Use when writing code that parses source, extracts structure/imports/exports/symbols/docstrings/comments, detects a language, runs diagnostics, or chunks code for LLMs — in Rust, Python, Node.js/TypeScript, or the ts-pack CLI. | | parsing-source | Use when the user wants a tree-sitter syntax tree for a source file — an s-expression dump or JSON tree. | | extracting-code-structure | Use when the user wants structured code metadata — functions, classes, imports, exports, symbols, docstrings, comments, or syntax diagnostics. | | chunking-for-llms | Use when splitting source code into chunks for an LLM context window without breaking syntax mid-construct. | | detecting-languages | Use when the user wants to know which programming language a file or snippet is — by path, extension, or content. | | managing-parsers | Use when managing the parser cache — prefetch for offline/CI, list downloaded languages, inspect a language, find the cache dir, or clean it. | | using-the-mcp-server | Use when parsing, processing, or detecting languages through the MCP server's parse/process/detect tools instead of the CLI. Covers the tool surface and the auto-installing launcher. |

MCP server

The plugin auto-registers an MCP server named tree-sitter-language-pack, launched via scripts/mcp-launch.sh (which execs ts-pack mcp). It exposes ts-pack's parse (syntax tree), process (structure/imports/exports/symbols/ docstrings/diagnostics/chunks), and detect (language detection) capabilities as tools, so the agent can parse and analyze code directly without shelling out to the CLI. The launcher auto-installs a binary on first run (override with TS_PACK_LAUNCHER=auto|npx|uvx|brew|download). The mcp subcommand ships in a recent release of the tool; an older binary on PATH may need an upgrade to expose it. See the using-the-mcp-server skill for details.

CLI

ts-pack subcommands:

| Command | Purpose | |---------|---------| | parse <file> | Parse into a syntax tree (--format sexp\|json). | | process <file> | Code-intelligence pipeline → JSON (feature flags below). | | list | List languages (--downloaded, --manifest, --filter). | | info <language> | Show whether a language is known and cached. | | download [langs...] | Download parsers (--all, --groups, --fresh). | | clean | Remove all cached parsers (--force). | | cache-dir | Print the cache directory. | | init | Write a language-pack.toml config. | | completions <shell> | Generate shell completions. |

process flags: --structure, --imports, --exports, --comments, --symbols, --docstrings, --diagnostics, --all, --chunk-size <bytes>, --language <name>. With no feature flags it defaults to --structure --imports --exports.

There is no detect or validate subcommand: language detection is implicit in parse/process (from the file extension), and diagnostics come from process --diagnostics. Content-based detection is available in the SDK (detect_language_from_content).

SDK

from tree_sitter_language_pack import process, ProcessConfig

result = process(source_code, ProcessConfig("python").all())
for item in result["structure"]:
    print(item["kind"], item["name"], item["start_line"])

import { process, detectLanguageFromPath } from "@kreuzberg/tree-sitter-language-pack";

const lang = detectLanguageFromPath("src/app.ts");
const result = process(source, { language: lang ?? "typescript", structure: true, imports: true });

Configuration

ts-pack init writes a language-pack.toml pinning a cache directory and a language set:

cache_dir = ".ts-pack-cache"
languages = ["python", "rust", "typescript"]

Commit it to pin the parser set across a team or CI. The cache location can also be inspected with ts-pack cache-dir.

Examples

Parse a file into a syntax tree:

ts-pack parse src/main.rs

Extract structure and imports as JSON:

ts-pack process src/app.ts --structure --imports

Chunk a large file for an LLM:

ts-pack process big_module.py --chunk-size 2000

Versioning

The plugin version tracks the marketplace VERSION file. See CHANGELOG.md for release notes.

License

MIT. The upstream tree-sitter-language-pack library is also MIT-licensed.

See also

• Marketplace: kreuzberg-dev/plugins
• Upstream: kreuzberg-dev/tree-sitter-language-pack
• Sibling plugins: kreuzberg, kreuzcrawl, kreuzberg-cloud