One Docker host
The simplest production setup: one Linux VM, the genosyn container, a managed Postgres next door. Behind your reverse proxy.
- Single-tenant, single-region
- Postgres on RDS / Cloud SQL / your own
- Caddy or nginx in front for TLS
The same open-source platform you can install in one command — only wired into your VPC, your Postgres, your identity provider, and your model keys. We help you ship it.
Genosyn was designed to be self-hostable on day one. For teams with real compliance gravity, running it in your environment is the short path — not the long way around.
Souls, run logs, attachments, customer records — all of it stays on infrastructure you control. No third-party data plane to vet.
Whatever your security team already audits — networking, KMS, IAM, logging — Genosyn inherits it. We help you map the controls.
Point Genosyn at your own Anthropic or OpenAI keys, or any OpenAI-compatible endpoint you host. Bedrock and Vertex are on the roadmap. No usage routing through a vendor.
Pre-pulled images, offline installer, optional self-hosted model gateways. Works on networks that never see the public internet.
Genosyn is a single container that talks to a database. That gives you a real menu of deployment options instead of one opinionated cloud product.
The simplest production setup: one Linux VM, the genosyn container, a managed Postgres next door. Behind your reverse proxy.
Run inside the AWS / GCP / Azure account that already holds your data. We help with IAM, KMS, security groups, and outbound model traffic.
Helm-managed deployment with horizontal replicas in front of a shared Postgres, an external volume for the data directory, and your cluster's secret manager.
Ship the image into your registry, point Genosyn at an internal model gateway, and run on networks the cloud never reaches. Updates come on your schedule.
Genosyn is MIT-licensed and you can absolutely run it on your own. For teams that want to skip the trial-and-error, we sell the shortest path from contract to first scheduled routine.
We pair with your platform team on the first deploy — provisioning, secrets, reverse proxy, the first AI employee — and hand back a runbook.
Help sizing Postgres, picking the right HA pattern, planning your backup and DR story, and migrating from SQLite when you outgrow it.
Threat-model walkthrough, hardening checklist, secret handling review, and answers for your security questionnaire from people who built it.
Wire Genosyn into Okta, Entra, Google Workspace, or any OIDC / SAML provider. Roles and provisioning kept in sync with your IdP.
Need an MCP server for an internal system, a private model gateway, or a one-off connector? We can build it with you or for you.
A shared channel, named on-call, and a response-time SLA that fits how mission-critical your AI employees actually are.
We built Genosyn to be the kind of thing your platform team can say yes to without a six-month review. One container, your database, your keys, your CI — and a vocabulary they already know.
Read the self-hosting docsSingle image, no sidecars, no microservice fan-out. Easy to audit, easy to roll back.
SQLite for dev, Postgres in prod. Same entities, same migrations, your backup story.
Claude, GPT, or any OpenAI-compatible endpoint you self-host. We never sit in the inference path.
Every line of what runs in your environment is on GitHub. Diff a release before shipping it.
Tell us where you want to run Genosyn, what compliance posture you need to meet, and how many AI employees you expect to staff. We'll come back with a deployment plan and a price.