A copilot for real data work
Not just another chatbot
No semantic layer required, no pre-cleaning required. Install the CLI, connect to your database and start querying. Bring your own API key and run it on your organization's approved model.
Does Your AI Analyst Solve the Right Problem?
Data analysts spend 80% of their time on dirty work. Most of AI tools on the market targets the other 20%.
“Dirty Work”
Actual Analysis
Writing the final query, building the dashboard, delivering the report.
Typical solutions: all trying to avoid the hard part — The REAL data work
You're Responsible for the Semantic Layer
Someone on your team must review every table and column, clean up and write descriptions. AI doesn't help with the hard part at all
Upload Clean Files
Export to Excel or CSV and upload to the platform. Some tools offer industry templates for accountants, finance, or HR — but either way, you're bringing pre-cleaned, structured data to their system.
Pick Your Top Tables
Got a database with thousands of tables? Give us your 10–15 most important ones. We'll help you write descriptions for those and build a chatbot on top — covering maybe 70% of your questions.
Zero Setup · Query Immediately
One Command Install
One command installs the CLI. No app to download, no environment to configure. Just install and you're ready.
Your Model, Your Rules
Use your preferred model or your company's approved LLM. Bring your own API key — no lock-in, no data sent to third parties.
Query on Day One
No semantic layer to build. No data cleaning required. Connect to your existing database and start querying immediately.
We NEVER see your data or queries
The CLI runs on your machine. Your conversation with the AI happens directly between you and your chosen LLM vendor. Your database connection happens directly between your machine and your database. We're not in the middle of either.
Your data and queries stay between your machine, your database, and your LLM vendor.
We only ship you the CLI tool and a license key. Your data never reaches our servers.
You choose your LLM vendor and accept their data terms (BAA included if you handle PHI).
Self-hosted Edition
Need air-gapped? Run dbology against your own LLM.
DBOLOGY can run against a locally-deployed LLM in your own environment. We can help with deployment for organizations that need full data isolation.
Introducing
Four commands we're proud of
Lineage and data-quality checks aren't new — doing them well on a real, messy warehouse is. These four dig into what metadata and dbt miss: stored procedures, dynamic SQL, notebooks, and the columns underneath.
/trace-upperWhere does this come from?
Lineage that doesn't stop at the catalog. It follows stored-procedure loaders, resolves dynamic SQL, reads notebooks and dbt models, and traces flow down to the column — reaching sources that metadata- or dbt-only tools never see.
/trace-downWhat breaks if I change this?
Know the real blast radius before you touch anything. It propagates impact column by column, then ranks what's actually used — by 30-day query volume and the jobs that hit it — so the consumers that matter surface first.
/dedupIs this column duplicated?
Goes beyond matching names. It compares how two columns are actually computed — logic, source, grain, time and tenant scope — walks back to where the value really originates, and tells you how confident it is.
/riskCan I trust these numbers?
Not just a null check. 33 checks across six dimensions — catching disguised nulls, sentinel values, and inverted-logic false alarms — each rated by confidence, so you know which numbers to trust.
One product · Many domains
dbology's general mode handles any database. Industry plugins layer in domain knowledge, terminology, public-data references, and ready-made calculations — so the AI reasons like a senior analyst in your field, not a generic SQL writer.
Healthcare Plugin
AvailableBuilt for analysts and engineers in payer, provider, and managed-care organizations.
Built-in healthcare expert
Ask HCC weights, HEDIS benchmarks, Medicare fees, ICD-10 / HCPCS / CPT — sourced from a curated reference catalog, not the model's memory.
Canonical healthcare data models
Built-in awareness of industry-standard schemas for claims, eligibility, and clinical data — the AI knows the structure of healthcare warehouses out of the box.
Speaks healthcare
Understands UM, MLR, Retro Amount, CAP Leakage, and other payer/provider jargon. Ships with pre-loaded memory so the first session already speaks the language.
Food & Beverage Plugin
Coming soonFor chain-restaurant and beverage operators — store comps, menu mix, daypart trends, and beverage SKU economics.
Your industry next?
Tell us your domain and what your team calculates every week.
Talk to usThe demo is ready
Two enterprise clients are already running it on their real databases. If you want to see it work on yours, we'd love to talk.