Open Source Context Graph

The context graph that
versions like git

Schema-as-code. Branch and merge. Local-first. Deploy when needed.

macos / linux

schema.pg

Branching

Branch your data like you branch code.

Every agent gets an isolated branch. Enrich, migrate, or remodel in parallel without touching main. A human reviews the results and merges what works.

Branch GraphOpen full view →

Isolation

Safe writes, zero contention

Each agent writes to its own copy-on-write branch. No locks, no conflicts, no risk to production data. Schema validation on every mutation.

Collaboration

Humans and agents on one graph

A human starts a branch, an agent picks it up. Or an agent enriches data and a human reviews it. The branch is the handoff.

Experimentation

Try it on a branch first

Schema migrations, bulk enrichments, new edge types. Test any change on a branch. If it breaks, discard. Main stays clean.

Shaping

data point → ontology → knowledge graph

From fragmented data to connected graph.

Features

One engine, the whole graph layer.

Git Workflows

Branches, merges,
transactional runs.

Git-style graph workflows. On your data, not your code.

Schema-as-code

$omnigraph schema plan ./repo

✓ 4 types compiled

✓ 11 of 11 queries typechecked

✓ lint clean

validation + linting at compile time.

Typed Schema

node Paper {
  doi: String @key
  title: String
  emb: Vector(768) @embed
}
node Claim {
  id: String @key
  text: String
}
edge Cites: Paper → Paper
edge Supports: Paper → Claim

Search + Traversal

traversalbm25vectorrrffuzzytext

Policy-as-code

Cedar for server-side access.

Typed schema, typed queries, typed mutations. No silent drift. No query surprises.

S3-native Storage

AWS S3RustFSON-premhybridVPC

Single CLI

initloadreadchangebranchmergesnapshot

Use Cases

Model any domain in a typed graph

Research Agents

Papers, notes, concepts, and claims connected in a typed graph. Agents trace citations, surface contradictions, and synthesize across sources.

Context Graphs

On-prem agent memory that survives sessions. Typed schema per agent, branches for experiments, Cedar policies at the edge.

Revenue Operations

Accounts, contacts, deals, signals, and activities in one typed graph. Agents score risk, surface gaps, and brief reps automatically.

Incident Response

Services, alerts, incidents, and actions in a dependency graph. Agents calculate blast radius, find similar past incidents, and propose mitigations.

Compliance

Policies, controls, evidence, and risks versioned in a graph. Agents flag gaps, trace decision chains, and surface overdue reviews.

Enterprise Knowledge

Multi-tenant, policy-gated knowledge for the whole org. Per-team branches, Cedar at the query layer, full audit.

Industry Intel

Companies, people, deals, filings, and patents connected over time. Agents detect signals, map actors, and brief analysts weekly.

Foundation

Open source all the way down.

No garbage collector. Memory safety without runtime overhead. Single binary, minimal dependencies.

In-memory columnar execution. No serialization overhead. The same runtime powering Pandas, DuckDB, and Polars.

Open format with built-in versioning. Each write is an immutable snapshot. Random access, not just column scans.

Production-grade query planner. Compiles typed graph queries to optimized execution plans.

Rust-native S3 backend. Same graph on laptop, VPC, or cloud — no config drift, no vendor lock-in.

Amazon's authorization language at the query layer. Deny-first, auditable access control.

Get Started

Bootstrap Omnigraph in 1 minute.

Omnigraph is open source, ships release binaries, and is ready to run.

GitHub Read the Docs

The context graph thatversions like git