ArcFlow
Company
Managed Services
Markets
  • News
  • LOG IN
  • GET STARTED

OZ brings Visual Intelligence to physical venues, a managed edge layer that lets real-world environments see, understand, and act in real time.

Talk to us

ArcFlow

  • World Models
  • Sensors

Managed Services

  • OZ VI Venue 1
  • Case Studies

Markets

  • Sports
  • Broadcasting
  • Robotics

Company

  • About
  • Technology
  • Careers
  • Contact

Ready to see it live?

Talk to the OZ team about deploying at your venues, from a single pilot match to a full regional rollout.

Schedule a deployment review

© 2026 OZ. All rights reserved.

LinkedIn
ArcFlow Docs
Start
  • Quickstart
  • Installation
  • Bindings
  • Platforms
  • Get Started
  • Cookbook
Concepts
  • World Model
  • Graph Model
  • Evidence Model
  • Observations
  • Confidence & Provenance
  • Proof Artifacts & Gates
  • SQL vs GQL
  • Graph Patterns
  • Parameters
  • Query Results
  • Persistence & WAL
  • Snapshot-Pinned Reads
  • Error Handling
  • Execution Models
  • Causal Edges
  • Adapter Discipline
  • Time Decay
  • Layers
  • 1. World Store
  • 1a. World Store · Smart Reader
  • 2. Perception Lake
  • 3. World Graph
  • 4. Query Engine
  • 5. Live Surface
  • 6. Event Bus
  • 7. Behavior Engine
  • 8. Algorithm Library
  • Virtual Computed Columns
  • Threading Model
  • Typed ID Contract
WorldCypher
  • Overview
  • Execution Options
  • Statements
  • MATCH
  • WHERE
  • RETURN
  • OPTIONAL MATCH
  • CREATE
  • SET
  • MERGE
  • DELETE
  • REMOVE
  • Composition
  • WITH
  • UNION
  • UNWIND
  • CASE
  • Schema
  • Schema Overview
  • Indexes
  • Constraints
  • Functions
  • Built-in Functions
  • Aggregations
  • Procedures
  • Shortest Path
  • EXPLAIN
  • PROFILE
  • Temporal Queriesfacet
  • Spatial Queriesfacet
  • Algorithmsfacet
  • Triggers
Capabilities
  • Live Queries
  • Vector Search
  • Trusted RAG
  • Spatial Knowledge
  • Temporal
  • Behavior Graphs
  • Graph Algorithms
  • Skills
  • CREATE SKILL
  • PROCESS NODE
  • REPROCESS EDGES
  • Sync
  • Programs
  • GPU Acceleration
  • Agent-Native
  • MCP Server
  • Event Sourcing
  • Intent Relay
  • Event Bus
Use Cases
  • Agent Tooling
  • Trusted RAG
  • Knowledge Management
  • Behavior Graphs
  • Autonomous Systems
  • Physical AI
  • Digital Twins
  • Robotics & Perception
  • Sports Analytics
  • Grounded Neural Objects
  • Fraud Detection
Walkthroughs
    Guides
  • Agent Integration
  • Building a World Model
  • Modeling a Social Graph
  • Build a RAG Pipeline
  • Using Skills
  • Behavior Graphs
  • Swarm & Multi-Agent
  • Fleet Coordination
  • Migrate from Cypher / Neo4j
  • From SQL to GQL
  • Filesystem Workspace
  • Data Quality
  • Code Intelligence
  • Scale Patterns
  • v0.7 → v0.8 Lakehouse Fast-Path
  • Tutorials
  • Knowledge Graph
  • Entity Linking
  • Vector Search
  • Graph Algorithms
  • Recipes
  • CRUD
  • Multi-MATCH
  • MERGE (Upsert)
  • Full-Text Search
  • Batch Projection
  • Multi-Source Observation
  • Sports Analytics
Operations
  • CLI
  • REPL Commands
  • Snapshot & Restore
  • Filesystem Projection
  • Plugin Management
  • Agent Governance
  • Server Modes & PG Wire
  • Persistence (ops)
  • Import & Export
  • Deployment
  • Deployment Modes
  • Daemon (UDS)
  • Why not Docker
  • Architecture
  • Engine Architecture
  • Cloud Architecture
  • Sync Protocol (Deep Dive)
  • World Graph Substrate (Preview)
Reference
  • TypeScript API
  • Glossary
  • Naming & Domain Map
  • Data Types
  • Operators
  • Error Codes
  • GQL Reference
  • Known Issues
  • Versioning
  • Licensing
  • Conformance
  • GQL Conformance
  • openCypher TCK
  • Extension Regressions
GQL Reference
    Conformance
  • Conformance Dashboard
  • openCypher TCK Results
  • Extension Regressions
  • Features
  • MATCH Basic
  • CREATE Nodes Edges
  • SET REMOVE Properties
  • DELETE Detach DELETE
  • RETURN WITH WHERE
  • Order BY Limit Skip
  • Order BY Nulls First Last
  • UNWIND
  • Aggregate Functions
  • OPTIONAL MATCH
  • Variable Length Paths
  • Label OR AND NOT Expressions
  • Label Wildcard
  • Quantified Path Sugar
  • Path Modes Walk Trail Simple Acyclic
  • Shortest Path Variants
  • IS Labeled Predicate
  • Element ID Function
  • IS Type Predicate
  • Binary Literals
  • Line Comments Solidus
  • Line Comments Minus
  • GQLSTATUS Result Codes
  • GQL Error Code Mapping
  • Transaction Control Syntax
  • SET Session
  • Conditional Execution WHEN THEN ELSE
  • RETURN NEXT Pipeline
  • Primary Key Constraint
  • Unique Constraint
  • Deterministic MERGE Via PK
  • Undirected Edge MATCH
  • Cast Type Conversion
  • GQL Directories
  • Multiple Labels Per Node
  • GQL Flagger
  • NEXT Linear Composition
  • Cardinality Function
  • INT64 BIGINT Type Names
  • FLOAT64 Double Type Names
  • Log10 Log2 Functions
  • Trim Leading Trailing Both
  • FILTER Clause
  • LET Statement
  • Group BY Explicit
  • EXCEPT SET Operations
  • INTERSECT SET Operations
  • ALL Different Predicate
  • Same Predicate
  • Property Exists Function
  • Path Variable Binding
  • USE Graph Clause
  • FOR IN List
  • Typed Temporal Literals
  • Session SET Value Params
  • Typed List Annotations
  • arcflow.cosine() function
  • arcflow.embed() function
  • arcflow.similar() procedure
  • arcflow.graphrag() procedure
  • ArcFlow Extensions
  • LIVE Queries
  • Triggered Write-Back Views
  • Evidence Algebra
  • Relationship Skills
  • AI Function Namespace
  • Graph Embedding Algorithms
  • ASOF JOIN
  • Durable Workflows
  • Incremental Z-Set Engine
  • GPU GraphBLAS
  • Triggers
  • HNSW Vector Index
  • Extensions Moat

Installation

Quick install (CLI today)#

The canonical install command on macOS + Linux:

curl -fsSL https://staging.oz.com/install/arcflow | sh

This is the only install URL the docs publish. If a stale CLI build surfaces a different URL (any other domain, including a deployment-host alias), prefer the canonical command above; the published install URL is the source of truth.

What this gives you today: the arcflow CLI binary at ~/.arcflow/bin/arcflow (or platform equivalent), and that's it — a single executable. The install message will say "Binaries: arcflow" (singular) — that is the current shipping scope.

What this does NOT give you today: the Python wheel, the npm package, or the Rust crate. Those are planned for RAM-C2 / 2026-Q3 on their respective public registries (the current shipping status for each binding lives in the install matrix below). Until they publish, the public-registry installs for oz-arcflow / arcflow will fail with No matching distribution. See Using ArcFlow from Python or Node today below for the recovery path.


ArcFlow ships across multiple registries. This page is generated from a single source of truth — the engine's RELEASE-MATRIX.toml. Every binding listed below was probed by CI before this page was published; planned artifacts are flagged with their target quarter and never advertised as if they were live.

Available now

  • Native CLI binary
    curl -fsSL https://staging.oz.com/install/arcflow | sh
    Platforms: darwin-arm64, linux-x86_64-gnu, linux-arm64-gnu, linux-x86_64-musl, linux-arm64-musl, windows-x86_64, windows-arm64
  • Standalone arcflow-daemon binary (UDS / named-pipe JSON-RPC)
    curl -fsSL https://github.com/ozinc/arcflow/releases/latest/download/arcflow-daemon-${VERSION}-${PLATFORM}.tar.gz | tar -xz
    Platforms: darwin-arm64, linux-x86_64-gnu, linux-arm64-gnu, linux-x86_64-musl, linux-arm64-musl, windows-x86_64, windows-arm64
  • ArcFlow MCP server (stdio JSON-RPC for AI agents)
    curl -fsSL https://github.com/ozinc/arcflow/releases/latest/download/arcflow-mcp-${VERSION}-${PLATFORM}.tar.gz | tar -xz
    Platforms: darwin-arm64, linux-x86_64-gnu, linux-arm64-gnu, linux-x86_64-musl, linux-arm64-musl, windows-x86_64, windows-arm64
  • libarcflow shared library (C ABI cdylib)
    curl -fsSL https://github.com/ozinc/arcflow/releases/latest/download/libarcflow-${VERSION}-${PLATFORM}.tar.gz | tar -xz
    Platforms: darwin-arm64, linux-x86_64-gnu, linux-arm64-gnu, windows-x86_64, windows-arm64

In progress

  • pip install oz-arcflow2026-Q3
  • npm install arcflow2026-Q3
  • cargo add arcflow2026-Q4

Manifest-driven. When a binding flips to shipped in RELEASE-MATRIX.toml, every install surface picks it up automatically.

By design — not provided

  • Docker image

    Refused 2026-04-30. ArcFlow is a 5MB statically-linked embedded library; Docker would add ~20MB of container overhead and subvert the in-process design. Customers should: - import the language binding (oz-arcflow Python wheel, arcflow npm package, arcflow Rust crate) for embedded use - use the native CLI for one-shot or scripting use - run the WASM playground at https://staging.oz.com/engine for the zero-install demo case If a future use case (long-running arcflow-mcp service, shared multi-tenant deployment) needs a containerized shape, add a separate `kind = "mcp-docker"` or `kind = "service-docker"` entry rather than reviving generic engine Docker.

ArcFlow ships as a 5MB embedded library — Browser, Node, Python, Rust, native CLI, MCP. Containerization would add overhead and subvert the in-process design. The engine is small enough to fit every deployment shape that earns its weight.

Upgrading from a pre-2026-05-14 install (the v1.x line)#

If you have an arcflow binary installed from before 2026-05-14 — anything reporting v1.x.y when you run arcflow --version — your arcflow upgrade command cannot migrate you to the current line. The v1.x line was retired 2026-05-14 when ArcFlow re-versioned to alpha (current line: v0.7.x and v0.8.x); legacy v1.x binaries hold their own retired upgrade-channel URLs in their compiled bytes and cannot self-correct.

Recovery: reinstall via the canonical install script.

curl -fsSL https://staging.oz.com/install/arcflow | sh

(See Quick install at the top of this page for the URL-stability context — the install URL standardized on staging.oz.com is what works today; the production-canonical oz.com/install/arcflow resolves once the oz-platform dev → production deploy lands, but the docs publish the working URL.)

After reinstall, verify:

arcflow --version    # should show v0.7.x or v0.8.x (current line)
arcflow upgrade      # works correctly from v0.7.x+; pulls the latest cut

If you're already on v0.7.x or v0.8.x: nothing to do. arcflow upgrade works correctly on the current line; you can keep using it.

Why this happens (one paragraph)#

ArcFlow moved from 1.x to 0.x versioning per an explicit operator directive — the alpha line restarts at 0.x, with v1.0 reserved for the first production-ready release. The arcflow upgrade code in any binary shipped before that pivot was compiled against the retired v1.x release channel; it has no awareness of the line move. The fix is to download a current binary via the install script (above), which then has the corrected upgrade-channel awareness for every future arcflow upgrade call. This is a one-time recovery per machine; no further intervention needed.

Using ArcFlow from Python or Node today#

If your project's pyproject.toml or package.json depends on oz-arcflow / @ozinc/arcflow and pip install / npm install returns "No matching distribution found", the cause is the publish gap shown in the matrix above — the SDK wheels are not yet on the public registries.

Until they ship, the CLI is the complete answer for most workflows. Install it with the curl one-liner at the top of this page and address ArcFlow through the JSON-over-stdin transport:

arcflow query "MATCH (n) RETURN count(*)" --json

The JSON-over-stdin transport is identical in shape to what the SDK wraps; everything you would call through the SDK is reachable from the CLI today. When the SDK wheels publish, switching from CLI shell-out to in-process calls is mechanical — the result envelopes are byte-identical.

How this page stays accurate#

The matrix above is rendered from release-matrix.json, published on every release. CI asserts on every commit that:

  • A status: shipped artifact's registry returns HTTP 200.
  • A status: planned artifact's registry returns HTTP 4xx (no false advertising).

If the manifest is unreachable at build time, the docs build fails — there is no fallback to a hand-rolled list. The matrix above is the live data; this page never ships an install command that the registry can't actually serve.

Verify installation#

import { openInMemory } from '@ozinc/arcflow'
 
const db = openInMemory()
db.mutate("CREATE (n:Hello {message: 'It works!'})")
const result = db.query("MATCH (n:Hello) RETURN n.message")
console.log(result.rows[0].get('message'))  // "It works!"
db.close()
from arcflow import ArcFlow
 
with ArcFlow() as db:
    db.execute("CREATE (n:Hello {message: 'It works!'})")
    for row in db.execute("MATCH (n:Hello) RETURN n.message"):
        print(row)  # {'message': 'It works!'}

The TypeScript example assumes the arcflow npm package is installed; the Python example assumes the arcflow Python wheel is installed. The matrix above shows you the install command for whichever path is currently shipped.

Environment variables#

VariableDefaultDescription
ARCFLOW_METAL_FORCE_UNAVAILABLEfalseSet to true to disable Metal GPU acceleration (useful if module load hangs)
ARCFLOW_NATIVE_DIRunsetPython only. Override the directory containing libarcflow.{so,dylib,dll} (used when running against a locally-built engine instead of the wheel-bundled binary).

Requirements#

  • Node.js 18+ (LTS recommended) for the npm path.
  • Python 3.8+ for the pip path.
  • No native build tools required — pre-built binaries are bundled in the installable artifacts.

Pre-release / nightly builds#

Per-commit pre-release artifacts are not published. The canonical release path is the GitHub Release on ozinc/arcflow — pin to a published tag for reproducibility.

For customer projects: pin oz-arcflow==X.Y.Z (Python) or @ozinc/arcflow@^X.Y.Z (npm) against the public registry. Don't vendor wheels — they go stale on every release, and the next teammate who clones the repo gets a binary that doesn't match the docs they're reading.

The vendored wheel rebuilds whenever the engine code changes; commit it alongside the test that depends on it for reproducible audit across sessions.


See Also#

  • Quickstart — first query in under 5 minutes
  • Project Setup — world model workspace layout and configuration
  • Language Bindings — Node.js, Python, Rust, Go, WASM, and CLI
  • Cookbook — runnable, manifest-aligned recipes
Try it
Open ↗⌘↵ to run
Loading engine…
← PreviousQuickstartNext →Bindings