The Relate Authors
Relate by Neo4j is an open-source, embedded graph database built in Rust for AI agents and agentic systems. It keeps connected facts and relationships close to the application, agent sandbox, notebook, browser, mobile or edge device, with no server to deploy.
At a Glance
- AI-agent and agentic-system developers
- Software engineers building coding agents and agent tools
- Developers needing an embedded graph database in Rust, Python, JavaScript, Java, Go, Swift or C
- Teams building desktop, mobile, edge, browser and offline applications
- +1 more
AI Tools by The Relate Authors
(1)Relate
Embedded Rust Graph Database
Discussions
No discussions yet
Be the first to start a discussion about The Relate Authors
Latest News
Relate v1.0.1 released with Python record conversion, improved Bolt error handling, JSONL output changes, security documentation and installer updates.
Neo4j press release: Relate released as a new in-memory database purpose-built for AI agents.
Neo4j published Meet Relate by Neo4j, describing the embedded graph database, agent skill, MCP, language clients, browser build and use cases.
Relate v1.0.0 released as the first stable release, with a 1.0 core cleanup and removal of CDC replication slots.
Products & Services
Rust implementation of an embedded, in-memory property-graph engine using Cypher; stores can be disposable in-memory graphs, directories or single .rl files with transactional durability through a write-ahead log and checkpoints.
Command-line shell for interactive and scripted queries, with an installable skill and agent-context reference for coding agents.
Embedding surfaces for Python (relate-db), Node.js (@relate/relate), Rust, Java, Go, Swift and C.
WebAssembly build for running the database in browser applications and notebooks.
Market Position
Relate is positioned as an agent-native, local alternative to deploying a client/server graph database for small, single-process workloads: it embeds in the application, starts quickly, works in memory or a local file, and combines relationship traversal with vector/full-text search. It remains compatible with the Neo4j ecosystem through Cypher, Bolt, HTTP and Aura/Neo4j subgraph pull/push, while its documented limitations (one process, one graph, RAM-bounded working set and no cluster) distinguish it from larger server or clustered graph platforms.
Leadership
Executive Team
Oskar Hane
Project maintainer / core contributor
GitHub profile identifies him as a Senior Staff Software Engineer at Neo4j. The Relate v1.0.0 and v1.0.1 release pages credit him with major release, installer, packaging and documentation work.
Michael Hunger
Project maintainer / core contributor
GitHub profile says he has worked on the open-source Neo4j graph database for 15 years, most recently in product innovation and developer-product strategy, and organized Neo4j Labs efforts. His GitHub handle jexp is listed among Relate v1.0.0 contributors.
Founding Story
Neo4j launched Relate to address the gap between heavyweight server databases and Python dictionaries or JSON files used by agents: agents need lightweight, fast storage that preserves relationships and supports graph reasoning. The initial vision was an agent-native, in-process database that starts in milliseconds, speaks Cypher, can run locally or in a browser, and can move data to and from Neo4j/Aura when a project grows.
Business Model
Revenue Model
Open-source software distributed under Apache License 2.0; the public site offers direct installation and local use without signup. No paid Relate service or subscription model is described in the reviewed sources.
Target Markets
- AI-agent and agentic-system developers
- Software engineers building coding agents and agent tools
- Developers needing an embedded graph database in Rust, Python, JavaScript, Java, Go, Swift or C
- Teams building desktop, mobile, edge, browser and offline applications
- Neo4j/Aura users needing local graph slices or embedded workflows
- Agent memory and connected context across sessions
- Agent coordination and shared task/dependency state
- Local agent sandboxes and throwaway graphs
- Codebase context, dependency analysis and security/staleness audits
- Embedded desktop, mobile and edge applications, including offline apps
- Notebooks, browser data visualization and WASM applications