Connect where it lives
For systems behind your firewall, a small connector runs inside your network. For cloud sources like BigQuery, we run the connector for you, so there is nothing to install.
Data fabric
Your data, kept in step.
No warehouse project first.
The Beye data fabric connects the systems you already run, brings their data into one place and keeps pulling in changes as the business runs. Your team builds on current data instead of waiting for a warehouse.
Yes, even the AS/400.
001 / Before and after
From the server closet
to the workspace.
Most teams already have the data. It is spread across an ERP, a WMS, a database nobody wants to touch and a few spreadsheets. The fabric brings it together and keeps it current, so everything you build in Beye starts from the same numbers.
- Start building without a data warehouse project first
- Ask questions of current data, not last month’s export
- Rerun the same tool next week and get next week’s numbers
- Keep the systems you have, including the old ones
- IBM AS/400 in the server closet
- SQL Server behind the firewall
- On-prem ERP installed in 2004
- WMS & TMS nobody wants to touch
- Spreadsheets only Linda understands
002 / How it works
Connect once.
Stay current.
The fabric is maintained continuously, not loaded once and forgotten.
Keep it current
Recurring, CDC-style updates pull in what changed. Each table runs on its own schedule, say hourly during business hours and less often overnight and at weekends.
Pick up where it left off
If a connection drops or the connector box restarts, the next run catches up on the changes that arrived in the meantime. Your source system stays the source of truth.
See that it is working
Every table shows its recent runs, schema changes and how far behind it is. If a run goes missing, you hear about it in the app, on mobile, by email or in your own Slack. Push notifications are coming soon.
Connects to
- IBM AS/400 (IBM i)
- IBM DB2
- SQL Server
- Postgres
- MySQL
- MongoDB
- NetSuite
- BigQuery
- File drops
003 / What it catches
Inserts, quiet edits and deletes.
Not just new rows.
Most simple syncs only notice new rows. The fabric catches the changes old systems make quietly, too.
New rows
Picked up on every run, in order, from where the last run finished.
Quiet edits
Old systems often change a row without touching a timestamp. A look-back window you choose compares row fingerprints to find those edits.
Deletes
Rows that disappear at the source disappear from the fabric too. The keys are compared from run to run.
No change column?
Some tables have nothing to track changes by. Those can be refreshed in full on a schedule instead.
004 / Gentle on the source
Built to be a polite guest
on your production database.
Whoever owns the AS/400 will ask whether this slows down production. Here is how it stays out of the way.
We wipe our feet, too.
Read queries only
The connector only reads. Give it a read-only database account and it has no way to change a thing.
Filters at the source
“Only 2025 onward” or “only plant 10” goes into the source query itself, so rows you do not want are never pulled.
Small, bounded reads
Tables are read a page at a time, never in one giant scan. Queries can be capped with a time limit, and are cancelled if the connector drops off.
Heavy checks, less often
The light check for new rows can run often. The heavier checks for edits and deletes run on their own, slower schedule, set per table.
005 / For your IT team
No ports to open.
Nothing to babysit.
The questions IT asks before anyone signs off, answered up front.
- Outbound only: no inbound firewall ports and no VPN for us
- One container on a small Linux box inside your network
- Updates itself and starts back up on its own after a reboot
- Data and credentials encrypted, in transit and at rest
- Credentials are scrubbed from error reports before they leave the box
- Works with a read-only database account
006 / Trust and history
Trust the copy.
Go back in time.
Proof it matches
We check the copy against the source with matching counts and totals, over old and recent ranges. We can show you it matches, number for number.
A dated history
Fully refreshed tables keep a dated history. See what inventory looked like last Tuesday, not just today.
File drops
Can’t install anything? Drop the export file, CSV or zipped, and processing starts on its own.
007 / Then put it to work
Current data is the start.
Here is what comes next.
Platform overview
Ask questions, run real Python and build tools your team reruns, all on fabric data.
Solutions
See what connected data looks like for distribution, manufacturing, retail and more.
Why Beye
How Beye compares with static reporting, copilots and generic AI tools.
008 / Frequently asked questions
Questions welcome.
Especially about the old systems.
For systems inside your network, one small connector. For cloud sources like BigQuery, no: we run the connector for you. And if installing anything is off the table, you can drop export files instead.
It is built not to. The connector only reads, applies your filters inside the source query, reads in small pages and runs its heavier checks less often than the light ones. You choose each table’s schedule, so busy hours can stay quiet.
No. The connector only makes outbound connections, so there are no inbound ports to open and no VPN to set up.
No. Beye brings your source data into its own fabric and keeps it current, so your team can start building without finishing a warehouse project. If you already have a warehouse, Beye can work with that too.
IBM AS/400 (IBM i), IBM DB2, SQL Server, Postgres, MySQL, MongoDB, NetSuite and BigQuery, plus file drops for exports. Not on the list? If we can connect to an AS/400 that is 30+ years old, we are not too worried about yours ❤️ Tell us what you run. Older or custom systems can take a little extra integration work.
It depends on the source and how each table is scheduled. Updates run on a recurring cadence that you can set per table, for example more often during business hours and less often overnight. It is recurring updates, not instant streaming.
The next run picks up from where the last one finished, including changes that arrived during the outage. A missed run raises an alert in the app, on mobile, by email or in your own Slack or webhook, instead of failing quietly. Push notifications are coming soon.
Each table shows its recent runs and whether they succeeded, any schema changes, and how far behind it is. We can also show you that the copy matches the source, number for number.
009 / Your next step
Bring the messy part.
Let’s put it to work.
Yes, even that spreadsheet.
Bring a report, calculation or recurring check you run today. We’ll show you how Beye connects the data behind it and turns it into a tool your team can rerun.