We're building the SDKs, APIs, documentation, examples, and tooling to make graph-powered applications and agentic AI systems fast to build and hard to break.
Python today via Bolt-compatible drivers: point an existing Neo4j client at AtherGraph. More languages coming.
Python-first: Agent, Tool, Skill, Flow. Declare an agent, get graph structure.
Reference docs for the query surface, retrieval APIs, and the agent runtime.
Cypher recipes for graph + vector + time-travel queries: copy, paste, adapt.
Triage, reconciliation, and document-pipeline agents as working starting points.
From first knowledge graph to GraphRAG with cited paths, step by step.
End-to-end reference apps: fraud paths, enterprise search, agent workflows.
A single binary, no cluster required: the whole engine on your laptop.
Run it on AWS, Google Cloud, Azure, or your own datacenter. Same engine everywhere.
Waitlist members get early SDK access, direct engineering contact, and a say in what ships first.