Installation

graphlens is published to PyPI as a small core library plus a set of language adapters and a CLI. You install only the pieces you need; the core never pulls in an adapter you did not ask for.

Requirements

Python 3.13 or higher
uv is recommended but pip works too.

Core library

The core contains the models, contracts, registry, exceptions, and utilities — everything except the language-specific parsing.

pip install graphlens

uv add graphlens

Core + a language adapter

Each adapter is exposed as an extra on the graphlens distribution:

pip install "graphlens[python]"        # Python adapter (Tree-sitter + ty)
pip install "graphlens[typescript]"    # TypeScript adapter (Tree-sitter + Compiler API)
pip install "graphlens[go]"            # Go adapter (Tree-sitter + gopls)
pip install "graphlens[rust]"          # Rust adapter (Tree-sitter + rust-analyzer)
pip install "graphlens[link]"          # cross-language linker (graphlens-link)
pip install "graphlens[all]"           # every adapter + the linker

The PHP adapter is Docker-only — it is not published to PyPI, so there is no graphlens[php] extra. Use the Docker image, which bundles PhpAdapter together with the intelephense binary.

With uv:

uv add "graphlens[python]"
uv add "graphlens[typescript]"

The CLI

graphlens-cli provides the graphlens command (analyze, query, visualize, neo4j). It declares its own extras so you can pick the adapters and exporters you need:

pip install "graphlens-cli[python]"        # CLI + Python adapter
pip install "graphlens-cli[typescript]"    # CLI + TypeScript adapter
pip install "graphlens-cli[neo4j]"         # CLI + Neo4j exporter dependency
pip install "graphlens-cli[all]"           # CLI + Python + TypeScript + Neo4j

To serve a graph to coding agents over MCP, install the dedicated graphlens-mcp server instead — it is built on top of this engine.

uv add "graphlens-cli[all]"

After installation the graphlens entry point is on your PATH:

graphlens --help

Docker (all adapters + toolchains pre-installed)

The published image bundles the CLI with every adapter and the toolchains their resolvers drive (ty, Node, Go + gopls, Rust + rust-analyzer, PHP + intelephense). This is the supported way to get the Go, Rust, and PHP adapters, which are not published to PyPI, and the easiest way to run graphlens in CI with no local setup. Mount your project at /workspace:

docker run --rm -v "$PWD:/workspace" ghcr.io/neko1313/graphlens \
    analyze /workspace --output /workspace/graph.json

The image is published to the GitHub Container Registry on each release (:latest plus :X.Y.Z and :X.Y version tags). See the Docker guide for details.

What gets installed where

Distribution	Import package	Provides
`graphlens`	`graphlens`	core models, contracts, registry, utils
`graphlens[python]`	`graphlens_python`	`PythonAdapter`, `TyResolver`
`graphlens[typescript]`	`graphlens_typescript`	`TypescriptAdapter`, `TsResolver`
`graphlens[go]`	`graphlens_go`	`GoAdapter`, `GoplsResolver`
`graphlens[rust]`	`graphlens_rust`	`RustAdapter`, `RustAnalyzerResolver`
Docker image only	`graphlens_php`	`PhpAdapter`, `IntelephenseResolver`
`graphlens[link]`	`graphlens_link`	`link_graph`, `LinkResult`
`graphlens-cli`	`graphlens_cli`	the `graphlens` CLI

Verify the install

from graphlens import adapter_registry
print(adapter_registry.available())   # e.g. ['python', 'typescript']

If your adapter shows up in that list, the entry point was discovered and you are ready to go. Continue with the Quick Start.

Requirements​

Core library​

Core + a language adapter​

The CLI​

Docker (all adapters + toolchains pre-installed)​

What gets installed where​

Verify the install​