Docker image

ghcr.io/neko1313/graphlens is the supported, batteries-included way to run graphlens — especially in CI. It bundles the CLI with every adapter and every toolchain their resolvers drive:

• ty for the Python resolver
• Node for the TypeScript resolver
• Go + gopls for the Go resolver
• Rust + rust-analyzer for the Rust resolver

It is also the only supported way to get the Go and Rust adapters, which are not published to PyPI.

Pull

docker pull ghcr.io/neko1313/graphlens:latest

The image is published to the GitHub Container Registry on each release:

Tag	Points at
:latest	the most recent release
:X.Y.Z	an exact version (e.g. :0.4.0)
:X.Y	the latest patch of a minor line (e.g. :0.4)

Pin to :X.Y.Z in CI for reproducible runs.

Run

The container's entry point is the graphlens CLI, so the arguments you pass are the CLI arguments. Mount your project at /workspace:

# Print stats
docker run --rm -v "$PWD:/workspace" ghcr.io/neko1313/graphlens \
    analyze /workspace

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

# Build the interactive viewer (headless)
docker run --rm -v "$PWD:/workspace" ghcr.io/neko1313/graphlens \
    visualize /workspace --no-open --output /workspace/graph.html

Write outputs under the mount

--output /workspace/graph.json writes the file back to your host through the volume mount. Writing anywhere outside /workspace keeps the result inside the ephemeral container, where it is lost when the container exits.

Selecting languages

Auto-detection runs every bundled adapter that recognizes the project. Narrow it with --lang:

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

Exporting to Neo4j from the container

Point --uri at a reachable database. From a container, localhost is the container itself — use host.docker.internal (Docker Desktop) or a service name on a shared network to reach a Neo4j running on the host:

docker run --rm -v "$PWD:/workspace" ghcr.io/neko1313/graphlens \
    neo4j /workspace \
    --uri bolt://host.docker.internal:7687 \
    --user neo4j --password secret

In CI

The same image powers the recommended GitHub Actions workflow — run the job inside the container and skip all language setup:

jobs:
  analyze:
    runs-on: ubuntu-latest
    container:
      image: ghcr.io/neko1313/graphlens:0.4.0
    steps:
      - uses: actions/checkout@v4
      - run: graphlens analyze . --strict --output graph.json

Local pre-push hook

Run analysis before every push without installing any toolchains (.git/hooks/pre-push):

#!/usr/bin/env bash
set -euo pipefail
docker run --rm -v "$PWD:/workspace" ghcr.io/neko1313/graphlens \
  analyze /workspace --strict