AWS Marketplace,Try Free with AWS
Skip to main content

DATABEND SOLUTIONS · USE CASE

AI & machine learning

Give AI a shared data foundation

Prepare analytical and evaluation datasets, enrich records through your model services, and connect agents to controlled SQL access with MCP.

Prepare, enrich, and explore in one workflow

Keep source data and model outputs together. External UDFs call your inference service; SQL Tasks publish reusable datasets.

Architecture: From Application Data to AI-Ready Datasets

Scroll horizontally to explore the full diagram.

Architecture: From Application Data to AI-Ready DatasetsBusiness records and AI application events load into raw tables. SQL and external UDFs prepare and enrich records using a model service in your own infrastructure. Tasks publish datasets for analytics, controlled MCP access, and evaluation workflows.DATA SOURCESDATA APPLICATIONSBusiness dataTables and recordsAI app eventsJSON and model payloadsLoad dataS3 stages / source syncPreserve source recordsraw_dataBusiness context + VARIANTStable identifiersSQL + external UDFPrepare selected recordsCall your model serviceTasksProcess new recordsStore enrichment resultsai_datasetsSource + model outputsReusable analytical tablesYour model serviceExternal inference endpointYour infrastructureProduct BIMCP accessEval datasetsSQL analysisSQL · External UDFs · Incremental TasksMCP: production read-only · Session sandbox

Connect your model service

Call custom inference endpoints through external UDFs and retain outputs with their source records.

Give agents controlled access

Databend MCP supports production read-only access and session-scoped sandbox writes.

Build reusable datasets

Prepare shared data for product analytics, model comparisons, and evaluation workflows.

Explore the details

Connect AI to useful data

Start with one application dataset or a custom enrichment workflow.

732 S 6TH ST, STE R, Las Vegas, NV 89101, USA
SOC 2 Type IIGDPR
© 2026 Databend Cloud. All Rights Reserved.