Open source

Self-hosted in an afternoon

No sales call, no sandbox tenant, no “book a demo.” Clone it, bring it up, and put a real request through the gate. The seed data ships a full governance scenario — AI systems, policies, a risk register, and demo users — so the dashboard is interesting from minute one.

# quickstart — Docker Compose with healthchecks

git clone https://github.com/anchorgate-ai/anchorgate.git

cd anchorgate

cp .env.example .env

make up

make seed

# then route a workload through the gate

export OPENAI_BASE_URL=http://localhost:8080/v1

# Kubernetes instead?

helm install anchorgate deploy/helm/anchorgate

AnchorProxy

OpenAI-compatible gateway — point your base URL here.

:8080

AnchorEngine

Control-plane API, OpenAPI docs at /docs.

:8000

AnchorUI

Governance dashboard with seeded demo logins.

:3000

v0.1 — early and moving fast. APIs may change; we say so rather than pretend otherwise.

What's in the box

A schema-first stack, developed in the open

The Decision Event Schema is the single source of truth — the Go, Python, and TypeScript types are generated from it and drift-checked in CI. Build your own tooling against the same records the dashboard reads.

Proxy: Go 1.22, distroless image, self-probing healthcheck
Engine: Python 3.11 · FastAPI · SQLAlchemy 2 (async)
UI: Next.js 14 · shadcn/ui · WebSocket live stream
Ledger: TimescaleDB on PostgreSQL 16, append-only hypertable
Hot path: Redis 7 — compiled policy bundles, token burns, cache
Schema: One JSON Schema → generated Pydantic, TypeScript, Go types

Running agents somewhere regulated?

We're onboarding a small group of design partners — platform and GRC teams with LLM or agentic workloads that will face an EU AI Act, ISO 42001, or SOC 2 conversation this year. You get direct input on the roadmap; we get your ugliest governance requirements.

Become a design partner

or email hello@anchorgate.ai