GenBI you can read, run, and keep.
Wren AI is open core. The context engine that answers your questions is open source, and every commercial plan runs that same engine, in our cloud or yours.
- Engine open source on GitHub
- Your MDL lives in your git
- Cloud, VPC, or air-gapped
One engine / open core, commercial platform
- GitHub stars
- 16K+
- Release cadence
- Weekly
- Discord members
- 1,700+
- Community growth
- 5× YoY
Cloud-locked platforms ask for trust. Open core earns it.
Every chatbot asks you to trust a closed platform. Six questions show the difference between asking for trust and earning it.
| Cloud-locked AI analytics | Open-core Wren AIOpen core | |
|---|---|---|
| Can you read the code? | Closed. You see outputs, not logic. | Open source on GitHub, end to end. |
| Where do your definitions live? | Inside the vendor's platform. | MDL files in your own git repo. |
| What happens if you leave? | Export data, rebuild everything. | Keep the engine and your models. |
| Where can it run? | The vendor's cloud only. | Cloud, your VPC, or air-gapped. |
| How do you evaluate? | Book a demo, wait for sales. | Run it locally today, free. |
| Who improves it? | One vendor's roadmap. | A core team shipping weekly, plus 1,700+ community members. |
12 platforms, 22 buyer factors, one guide.
We scored open-core Wren AI against cloud-locked AI analytics platforms across 22 buyer-grade factors. The full tally is free to download and circulate with your team.
LLMs don't know your data. The open core does.
The open-source engine sits between agents and your data, enforcing shared definitions and metric logic at query time. Built for individual developers and used through the CLI, it's the layer you can read on GitHub.
- Grounded text-to-SQL, no schema-guessing
- MDL: metrics and definitions as code, in your git
- CLI & MCP server for your own agents, no UI required
- 20+ data sources, 60+ AI agents, any LLM
Connect once / 20+ data sources, queried in place
What's open, what's commercial. Published, not fine print.
Everything commercial stacks on the open core, and the boundary between them is published. The bottom layer is the one you have to trust, which is why it is the one you can read.
For individual developers, used through the CLI and MCP server (no UI). The part you have to trust is the part you can read, and this engine runs without us.
The engine stays open because the engine is what earns your trust.
Start where your constraints are.
Every path runs the same context engine. Models you build on one carry to the next, so you can move without rebuilding.
Wren AI OSS
For individual developers building their own agents. The engine and CLI, no UI.
- Free for individual developers
- Context engine, MDL, MCP server, and CLI (no UI)
- Community support on Discord
Wren AI Cloud
For teams: the full GenBI platform, managed by us. Connect a database and ask your first question today.
- Starts free, no infrastructure
- GenBI Apps, agentic mode, and dashboards
- Upgrade as your team grows
Enterprise Plus
For enterprises: the full platform inside your network, air-gap ready, with unlimited users.
- Runs in your VPC or on-premise
- Access control, row & column-level security
- Licensed by concurrent sessions, not seats
The questions skeptics should ask.
Read the code. Star the repo.
If open core is the model you want more of in your data stack, a star is how you say so. Run the engine locally today, and talk to us when you want the platform on top.