Choose where your data actually lives.
Four ways to deploy AI2BI Hub — same product, different boundary. Switch any time. Your AI credits and saved queries stay with your account, not with the mode.
Cloud
DefaultData stored in the AI2BI cloud (S3, encrypted at rest). Every feature enabled.
Included in every account — no extra setup needed.
What you get
- Files stored in AWS S3
Encrypted at rest with AWS-managed keys. Customer-Managed Keys (CMEK) available on Pro+ plans.
- Works across devices
Sign in from any browser or device and see the same files + saved queries.
- All features enabled
Chat, Auto-dashboard, Anomaly scan, NL chart builder, Schema editor — everything.
- Fastest startup
No local engine to boot. First query hits Athena or a warm Lambda in ~1–2s.
Heads up
- Data travels to our cloud
Raw rows are stored in S3 and queried server-side. If that’s a concern for your data, see Browser-Private.
Browser-Private
No installFiles live in this browser tab only. Raw rows never leave your machine.
Switch any time from Settings → Privacy mode.
What you get
- Files stay in this browser on your laptop
Your CSVs are saved inside this browser tab. They are not uploaded to our servers and are not synced to other devices.
- No install required
A small analytics engine loads into the browser the first time you use the mode (~8 MB). After that it runs instantly on every visit.
- AI billing unchanged
Chat still spends tokens. Only your question + column names go to our servers for the AI reply — the actual rows never leave your machine.
Heads up
- Anomaly scan disabled
It needs server-side access to your rows, which is the opposite of this mode. Dashboards, manual charts, and chat all work fine.
- Tied to this device + browser
Your data doesn’t follow you if you sign in from another laptop or clear your browser storage.
Enterprise Agent
On-premises agent runs on your Windows / Mac / Linux machine. Nothing leaves your device.
We’ll email you when on-premises rollout opens up.
What you get
- Installer runs locally
Small signed agent runs as a background service. Data files live wherever you choose on your disk.
- No cloud dependency for data
Only a periodic subscription heartbeat leaves your network. Every query, every LLM call runs against the local agent.
- Enterprise features
SSO (SAML / OIDC), custom retention, SOC 2 evidence pack.
Heads up
- Currently in private preview
Register for early bird below and we’ll reach out with a timeline.
MCP Server
Use AI2BI Hub directly from Claude Desktop, ChatGPT, or any MCP-aware AI tool. Your data stays where it is — the AI client connects to AI2BI as a tool.
Be first in line when the MCP server opens up.
What you get
- No new UI to learn
Ask questions inside Claude Desktop / ChatGPT / Cursor. The MCP server exposes "ask", "query", "list_files" etc. as tools the AI can call on your behalf.
- Same backend, new front door
Auth via your AI2BI session, results streamed back into the LLM client. Your existing tier + credit balance applies.
- Healthcare path requires Claude Enterprise
For T4 (PHI) tenants, MCP needs admin opt-in plus your own BAA with Anthropic / OpenAI Enterprise. We don’t enable it by default for healthcare accounts.
Heads up
- Currently in roadmap
We’re building this once Phase D (upload scanner) is rolled to all products. Watch the Trust page for the launch date.
Try any mode, free for 14 days
200,000 trial tokens. Switch between modes any time from Settings → Privacy mode. No credit card required.