Connect everywhere your business actually lives.
Not just your databases and warehouses — the streams, the SaaS tools, and the rules hiding in your data, your docs, and your code. Read in place with no copy, classified and governed on arrival, landed in open Iceberg on a bucket you own.
Four things that make connecting to Eisberg different.
Read in place — no copy required
Point Eisberg at PostgreSQL, SQL Server, Oracle, or Snowflake and it reads them where they sit, through engine federation. It classifies sensitive data at the source, masks per role using the source's own policy, and resolves the same real-world customer across owned and federated systems — with zero copy. You don't have to move your data to get value from it.
The rules nobody ever wrote down
Most of your real business logic isn't in a table — and most of it isn't written down anywhere at all. It's implicit: the vendor that's always excluded from aging, the 90-day cutoff baked into how DSO is really calculated, the segment that's quietly treated differently. Eisberg infers these rules by watching your data and how your teams actually query and act on it — the patterns, the exceptions, the anomalies — then corroborates them against what is written, in Slack, Confluence, GitHub, and stored procedures. Each inferred rule is bound to your business model as a governed fact, with its evidence and a confidence score, and surfaced for a human to confirm. The rules in people's heads become rules the platform can enforce.
Lands in open Iceberg — on your bucket
Whatever it ingests lands as Apache Iceberg in your own object storage (S3, Azure Blob, GCS) — never a proprietary vault. Any engine reads it: Spark, your existing warehouse, anything Iceberg-compatible. Your continuity is in the open format, not in Eisberg. If you ever leave, your data is already portable.
Governed the moment it arrives
Every source is classified on contact — PII, PHI, and sensitive fields tagged automatically — and masked by role from the first query. Every connection, every read, every action lands in a tamper-evident audit trail. Connecting a source turns governance on, it doesn't defer it.
What it connects to today.
Grounded in the connectors that ship — structured systems and the tribal-knowledge sources that back the whole “learns your business” thesis. Don't see yours? Most enterprise systems are a connector away, and anything with a REST or JDBC surface is already reachable.
Databases & warehouses — read in place, no copy
Streaming & change data capture
Files & object storage — yours, not ours
Business & SaaS systems
Tribal knowledge — comms, docs & code
Any API
AI providers — route to any model
Open standards we ship on
Publish & act to
Connected to useful in three steps.
Point it at a source
Give it a connection — a database, a bucket, an event-stream topic, a SaaS account, a Slack workspace. No pipeline to hand-build.
It reads, classifies, and learns
It reads in place or ingests to Iceberg, classifies sensitive data, and starts folding the source into your business model.
Everyone can use it — governed
The moment it's connected, any team can ask and act on it in plain language, masked by role, logged end to end.
Bring the source that matters most.
Pick the system you'd want connected first — a production database, a data lake, a Slack workspace full of tribal knowledge. We'll connect it live and show you Eisberg learning from it, governed the whole way.