Introduction
Welcome to Grengin. It’s an open-source, self-hosted AI workspace that gives your whole team governed access to the best large language models (LLMs), running in your own cloud.
Grengin brings together a branded chat interface, multi-provider model access, and an enterprise-grade governance layer (administration, SSO, cost controls, PII detection, and audit trails) in a single platform you can stand up in about five minutes. You bring your own provider API keys, your conversations stay in your own tenant, and Grengin is never in your data path.
AI for all. Control for IT. Insights for executives. In about five minutes.
What is Grengin?
Most teams that want to adopt AI safely are stuck choosing between options that don’t quite fit:
- Enterprise plans like ChatGPT Enterprise or Claude Enterprise carry large seat minimums and price tags built for big organizations.
- Microsoft 365 Copilot assumes you’re already a Microsoft shop.
- Free consumer tools send your prompts somewhere you don’t control, with no governance or audit trail.
- DIY self-hosting with projects like LibreChat gives you control but demands DevOps expertise most teams don’t have in-house.
Grengin is the middle path. It gives small and mid-sized teams the governance, security, and cost visibility usually reserved for enterprise deals, with the simplicity and pricing of a startup tool and none of the lock-in. Whether you’re a 10-person startup or a 500-person company, the platform scales with you.
The web app and API are fully open source and public on GitHub (grengin-oss/grengin), so your team can read every line, self-host it, and verify there’s no hidden behavior.
Why Grengin?
A few principles set Grengin apart from both enterprise suites and consumer chat tools:
Your keys, your data. You connect your own provider API keys, and prompts travel directly from your deployment to your AI provider. When you self-host, Grengin is never in the data path and never sees your prompts, completions, or logs.
No per-seat fees. There are no seats, no user minimums, and no plans to choose between. The software is free to build from source, and the one-click Marketplace image is metered only for the compute it uses. Cost scales with what you actually consume, not with headcount.
Governance from day one. SSO, centralized user management, per-department and per-project policies, budget caps, PII detection, and complete audit trails are built in, not bolted on through a separate enterprise contract.
No lock-in. Your deployment runs in your own cloud, your conversations live in your own tenant, and model access runs on your own keys. Switch models or providers whenever you like. If you ever leave, you keep your data and your keys.
Key features
| Feature | What it does |
|---|---|
| Multi-LLM access | Connect OpenAI, Anthropic (Claude), Google Gemini, and a range of open-source models: 7+ across providers, switchable per task, all under one governance layer. |
| Governance and administration | An admin dashboard with SSO, centralized user management, and policies scoped by department or project. |
| Cost control | Real-time cost tracking with budget alerts and hard caps per user, department, or project. |
| Security and oversight | PII detection that helps prevent sensitive data from leaving, plus complete, exportable audit trails of every interaction. |
| Integrations | Jira, Confluence, and Google Drive out of the box, plus anything that speaks MCP, the open standard for connecting AI to your tools. |
| Fast deployment | Live in about five minutes. No IT project, no consultants, no DevOps team required. |
How it works
Getting your team up and running takes three steps:
- Deploy Grengin into your own cloud: one click from the AWS or Azure Marketplace, or build from source for full control.
- Set up by signing in with SSO, connecting your model provider keys, and setting department policies and budgets.
- Invite your team over SSO. Governance, cost controls, and audit trails are on from day one.
From there, your team uses AI through the Grengin web app or the REST API.
Ways to run Grengin
Grengin is self-host-first, with three deployment paths that all run the same software:
| Path | Best for | Cost |
|---|---|---|
| Build from source | Teams who want full control and don’t mind a hands-on setup | Free. You pay only your cloud and AI providers |
| Marketplace image (AWS / Azure) | Most teams: a one-click VM deploy into your own cloud account | Metered at about $0.001 per vCPU-hour on your own cloud bill (roughly $3/month on a typical 4-vCPU VM) |
| Grengin Cloud (managed) | Teams that prefer not to run any infrastructure | Managed and Grengin-hosted |
On the self-hosted paths (source and Marketplace), your deployment is single-tenant inside your own AWS or Azure account, and Grengin stays out of your data path. Grengin Cloud is the managed option for teams that would rather not operate infrastructure; in that case, traffic flows through Grengin’s hosted environment.
Need data residency? On a self-hosted deployment you choose the cloud region, and your data never leaves your tenant.
What’s next?
Ready to get started? Head to the Quick Start guide to deploy Grengin and send your first prompt in minutes.
From here you can also:
- Compare the source, Marketplace, and managed paths in Deployment.
- Set up SSO, policies, and budgets in Configuration.