Skip to main content

Neo4j CLI

Project description

Neo4j CLI

Installation

Pick one. Verify with neo4j-cli --help.

  • Install script: curl -sSfL https://neo4j.sh/install.sh | bash.
  • Homebrew: brew install neo4j-labs/tap/neo4j-cli (stable releases only; prereleases ship via npm/PyPI).
  • Prebuilt archive: grab your OS/arch from releases.
  • npm: npm i -g @neo4j-labs/cli (also works with pnpm add -g / yarn global add). Prereleases: @alpha, @beta, @rc, @next. Platform matrix: distribution/npm/cli/README.md.
  • PyPI: pip install neo4j-cli, pipx install neo4j-cli, or uv tool install neo4j-cli. One-shot: uvx -i neo4j-cli <commands>. Pin a prerelease with ==, e.g. pipx install neo4j-cli==0.1.0a6.

Agent skills

neo4j-cli ships an embedded skill bundle (SKILL.md + per-subcommand references) that teaches AI coding agents how to drive the CLI. skill install drops it into each detected agent's skill directory; pass an agent name to target one.

Supported agents: Claude Code, Cursor, Windsurf, Copilot, Gemini CLI, Cline, Codex, Pi, OpenCode, Junie.

neo4j-cli skill install                  # all detected agents
neo4j-cli skill install claude-code      # one agent
neo4j-cli skill list                     # per-agent install state
neo4j-cli skill check                    # version drift vs running binary
neo4j-cli skill remove [agent]           # idempotent

Credentials

neo4j-cli stores three kinds of credentials in credentials.json under your OS config directory. All three trees share the same add / list / use / remove shape; use sets the default consumed by downstream commands.

credential aura-client — Aura Console API credentials (client ID + secret). Required for any aura ... subcommand that calls the Console API.

neo4j-cli credential aura-client add --name "my-org" --client-id <id> --client-secret <secret>
neo4j-cli credential aura-client list
neo4j-cli credential aura-client use my-org
neo4j-cli credential aura-client remove my-org

credential dbms — Neo4j Bolt connection profiles (URI, username, password, database, optional embed-credential link). When a default profile exists, neo4j-cli query connects without any connection flags or env vars.

neo4j-cli credential dbms add --name prod --uri neo4j+s://example.databases.neo4j.io --username neo4j --password '<pw>'
neo4j-cli credential dbms list
neo4j-cli credential dbms use prod
neo4j-cli credential dbms set-embed prod openai-shared    # link an embed credential
neo4j-cli credential dbms set-embed prod                  # clear the link
neo4j-cli credential dbms remove prod

credential embed — Embedding-provider credentials (provider, model, base URL, dimensions, optional API key). Consumed by query --param NAME:embed=... and query :embed. Supported providers: openai, ollama, huggingface.

neo4j-cli credential embed add --name openai-shared --provider openai --model text-embedding-3-small --api-key '<key>'
neo4j-cli credential embed list                           # api-key is never printed
neo4j-cli credential embed use openai-shared
neo4j-cli credential embed remove openai-shared

Aura

Manage Neo4j Aura instances from the terminal. Requires an aura-client credential — create one in your Aura Account Settings and add it via Credentials above.

List your instances

neo4j-cli aura instance list --format table

Create an instance

# Free-db — no cloud provider, region, or memory required
neo4j-cli aura instance create --name my-free-db --type free-db --tenant-id <tenant-id> --rw

# Professional-db on AWS, awaiting readiness
neo4j-cli aura instance create --name my-pro-db --type professional-db --cloud-provider aws \
  --region us-east-1 --memory 4GB --tenant-id <tenant-id> --await --rw

aura tenant list shows tenant IDs. Initial DB credentials returned by instance create are auto-stored as a dbms credential (named <instance-id>-default), so neo4j-cli query can connect immediately. Use --no-credential-storage to skip that.

Querying Neo4j

neo4j-cli query runs Cypher against any Neo4j database via the Bolt protocol. Cypher comes from the positional argument or piped stdin.

neo4j-cli query 'RETURN 1 AS n'
echo 'MATCH (n) RETURN count(n)' | neo4j-cli query

Preferred: add a dbms credential (see Credentials) and query connects with no further config. aura instance create auto-stores one for new instances.

Flags, env vars, and .env files are optional overrides — useful for one-offs or CI without persisting a credential. When a stored credential exists, an override must supply all four of URI/username/password/database (any partial set is rejected). Without a stored credential, resolution is flag → env var → .env file (auto-discovered walking up from cwd) → built-in default.

.env discovery walks up from cwd and stops at the first .git ancestor or your $HOME boundary (whichever comes first), so a .env outside your repo or above your home directory is never loaded. When the loaded .env lives strictly above cwd, an info: loading .env from <path> line is printed to stderr so the overlay is never silent.

Setting Flag Env var Default
URI --uri NEO4J_URI neo4j://localhost:7687
Username --username NEO4J_USERNAME neo4j
Password --password NEO4J_PASSWORD prompted on TTY
Database --database NEO4J_DATABASE neo4j

http:// and https:// URIs are auto-rewritten to neo4j://<host>:7687 and neo4j+s://<host>:7687 respectively (path/query stripped). For self-signed certs use neo4j+ssc://.

Pass parameters with --param key=value (repeatable). Values that parse as JSON are typed; everything else is a string:

neo4j-cli query 'MATCH (p:Person {name:$name}) RETURN p' --param name=Alice
neo4j-cli query 'RETURN $ids' --param 'ids=[1,2,3]'
echo 'MATCH (p:Person {name:$name}) RETURN p' | neo4j-cli query --param name=Alice

Output is a table by default; pass --format json for a stable envelope (columns, rows, truncated, arrays_truncated). When stdout is not a terminal (piped or redirected), --format defaults to json. Applies to both query and :schema. Large results are capped at 100 rows and arrays inside cells at 100 items — tune with --max-rows / --truncate-arrays-over (0 = unlimited).

Schema introspection:

neo4j-cli query :schema

Embedding parameters

Bind a vector parameter inline by passing --param NAME:embed=<text> — the text is sent to the configured embedding provider and the resulting []float32 is bound to $NAME for both the EXPLAIN preflight and the real run. The sibling query :embed [text] leaf computes a vector standalone without opening a Bolt connection.

neo4j-cli query --param q:embed='sci-fi movies' --param k=5 \
  "CALL db.index.vector.queryNodes('idx', \$k, \$q) YIELD node, score RETURN node, score"

neo4j-cli query :embed "hello world" --format json
echo "hello world" | neo4j-cli query :embed --format toon

Embedding settings resolve with this precedence (highest first): flag → env var → .env file → stored embed credential → provider built-in default.

Setting Flag Env var
Credential --embed-credential
Provider --embed-provider NEO4J_EMBED_PROVIDER
Model --embed-model NEO4J_EMBED_MODEL
Base URL --embed-base-url NEO4J_EMBED_BASE_URL
Dimensions --embed-dimensions NEO4J_EMBED_DIMENSIONS
API key (none — see below) NEO4J_EMBED_API_KEY, OPENAI_API_KEY, HF_TOKEN

API-key precedence (highest first): per-provider OS env (OPENAI_API_KEY / HF_TOKEN) → generic OS env (NEO4J_EMBED_API_KEY) → per-provider .env value → generic .env value → stored credential api-key. Ollama needs no API key.

--embed-credential <name> selects a stored embed credential explicitly; without it the resolver falls back to the embed credential linked from the resolved dbms credential (via credential dbms add --embed-credential or credential dbms set-embed), then to credential embed's default. So one --credential <name> can drive both DB connection and embedding when the dbms credential carries an embed link.

Provider defaults: OpenAI base URL https://api.openai.com/v1, Ollama http://localhost:11434, HuggingFace https://router.huggingface.co/hf-inference/models (serverless mode). Setting --embed-base-url switches HuggingFace to dedicated-endpoint mode.

Write operations

Write commands require --rw. neo4j-cli query runs EXPLAIN first when --rw is absent and blocks mutating Cypher before execution.

neo4j-cli aura instance delete <id> --rw
neo4j-cli config set telemetry false --rw
neo4j-cli query 'CREATE (:Person {name:"Alice"})' --rw

Feedback / Issues

Please use GitHub issues to provide feedback and report any issues that you have encountered.

Building locally

Clone the repository and run:

make build

This produces bin/neo4j-cli. To run without building:

make run-neo4j

Developing and contributing

Read CONTRIBUTING.md

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

neo4j_cli-0.1.0a9-py3-none-win_arm64.whl (13.5 MB view details)

Uploaded Python 3Windows ARM64

neo4j_cli-0.1.0a9-py3-none-win_amd64.whl (14.7 MB view details)

Uploaded Python 3Windows x86-64

neo4j_cli-0.1.0a9-py3-none-manylinux_2_17_x86_64.whl (14.3 MB view details)

Uploaded Python 3manylinux: glibc 2.17+ x86-64

neo4j_cli-0.1.0a9-py3-none-manylinux_2_17_aarch64.whl (13.3 MB view details)

Uploaded Python 3manylinux: glibc 2.17+ ARM64

neo4j_cli-0.1.0a9-py3-none-macosx_11_0_arm64.whl (13.9 MB view details)

Uploaded Python 3macOS 11.0+ ARM64

neo4j_cli-0.1.0a9-py3-none-macosx_10_9_x86_64.whl (14.8 MB view details)

Uploaded Python 3macOS 10.9+ x86-64

File details

Details for the file neo4j_cli-0.1.0a9-py3-none-win_arm64.whl.

File metadata

  • Download URL: neo4j_cli-0.1.0a9-py3-none-win_arm64.whl
  • Upload date:
  • Size: 13.5 MB
  • Tags: Python 3, Windows ARM64
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for neo4j_cli-0.1.0a9-py3-none-win_arm64.whl
Algorithm Hash digest
SHA256 8af2ee841c4ba7dd52f6b878b9f41b3b821013bdba64f3992e2b390f30af9e2c
MD5 915ff5ee553354af6fc2daeaa235a9a0
BLAKE2b-256 0d5e2d2a3637dbe5036460043fb349011ecd36e6cf789fbde6e386cfd0967d19

See more details on using hashes here.

File details

Details for the file neo4j_cli-0.1.0a9-py3-none-win_amd64.whl.

File metadata

  • Download URL: neo4j_cli-0.1.0a9-py3-none-win_amd64.whl
  • Upload date:
  • Size: 14.7 MB
  • Tags: Python 3, Windows x86-64
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for neo4j_cli-0.1.0a9-py3-none-win_amd64.whl
Algorithm Hash digest
SHA256 a9cf47eb1c981b3c7881913f5d0ea147f676e63240fa95799abbad8c4b2419f5
MD5 fff181e1e11fff9112fbd291da4a6a07
BLAKE2b-256 04721fdf36b6d973fb03f5f9b1349a1a6e07e123aae3fddeb533bc964d75b49b

See more details on using hashes here.

File details

Details for the file neo4j_cli-0.1.0a9-py3-none-manylinux_2_17_x86_64.whl.

File metadata

File hashes

Hashes for neo4j_cli-0.1.0a9-py3-none-manylinux_2_17_x86_64.whl
Algorithm Hash digest
SHA256 853908fb6d602bb004d88f31794e75b63fbc287f1f1a7606b14fb42dc721e1f4
MD5 71628207896eb6c2394cfbea2e086ee7
BLAKE2b-256 287dcd09546056321c9b7846a3fc1f42bd24b1a391ece96d90a4c90b1ce39bb4

See more details on using hashes here.

File details

Details for the file neo4j_cli-0.1.0a9-py3-none-manylinux_2_17_aarch64.whl.

File metadata

File hashes

Hashes for neo4j_cli-0.1.0a9-py3-none-manylinux_2_17_aarch64.whl
Algorithm Hash digest
SHA256 dfb82790ec473f6cb3aa712bca5307ba0f6a6d2cafd8ee39a94e5ab697b3551a
MD5 e2fe336cf62e9253018192952a98c95b
BLAKE2b-256 9b2f1a6eaad17d7068233d530732e070bfb8c348b14b29a27b9d6040e5fbb41a

See more details on using hashes here.

File details

Details for the file neo4j_cli-0.1.0a9-py3-none-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for neo4j_cli-0.1.0a9-py3-none-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 d45ffc3cb730e0e57d6ee308a341976f3e1309380b80e72e54f6c539ac7faefd
MD5 f583650faa2c6ada1ababb2c350c4aad
BLAKE2b-256 4c9a70b9628c80385ce301f6a0365bcb6ee33a88cf447c24684bf3ff04c4b7ed

See more details on using hashes here.

File details

Details for the file neo4j_cli-0.1.0a9-py3-none-macosx_10_9_x86_64.whl.

File metadata

File hashes

Hashes for neo4j_cli-0.1.0a9-py3-none-macosx_10_9_x86_64.whl
Algorithm Hash digest
SHA256 6bf08e8e98f97a1893e316c7d80e1443dd60167b5e9fac0e24417c400a1defa3
MD5 0f3051e7bdf5fecdfed97abfe1ce0227
BLAKE2b-256 ebe02f1c9205b03a62545d12b6fc99a5c894240e2dd91305db21f81190141913

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page