News & insights
Product updates and technical writing from Databasin: new capabilities, practical guides, and the engineering details behind the work.
Showing 36 entries.
Skills are now a first-class object rather than an admin-owned prompt blob. Read a skill's full instructions before you run it, name your own copy, edit it right where you use it, and choose who else gets it — this automation, this project, or your whole organization.
The biggest change to how data gets into Databasin since the connector catalog itself. Connect a supported source and it appears as live tables instantly, so you can SELECT straight against HubSpot, Sage, or your database while one-button sync keeps a lakehouse copy for the heavy lifting.
Starting Databasin is now a single field. Type your email, click once, and your sign-in link lands in your inbox with a workspace already built: $50 in credit, no card, no form. Sign in with Google or Microsoft and ask your first question.
Query one Apache Iceberg table from Trino, Spark, and DuckDB. See the same table through three engines, with setup and trade-offs.
The same AI that answers your questions can now work without you: a new agent task runs a skill — exec summary, data-quality check, anomaly explanation — inside your scheduled automations, with read-only tools, query budgets, and a full audit trail on every run.
A quality-of-life drop across the platform: global filters on dashboards, a Connection Credentials viewer for plugging external tools into any engine, Apache Doris streaming in notebooks, live IP whitelists in cluster dialogs, and a sharper SQL editor.
Spotify's Megaphone platform joins the native connector catalog: your full podcast catalog and every episode's metadata and publishing state, synced into the lakehouse across all your networks — ready for Databasin One to answer questions about your shows.
Everything in the in-app Learn Center — 39 guides across connectors, pipelines, the lakehouse, Databasin One, and administration — is now published at databasin.ai/docs, searchable and readable before you ever create an account.
A new Add Data wizard walks you from "I have data in X" to a running pipeline; Managed Connect lets us provision it for you with live progress; Connector Forge turns any OpenAPI spec into a connector; and every connector card now shows honest cost and time estimates.
Point a task at your dbt Core repo and Databasin does the rest: clone, generate profiles, run, and report per-model status back into the automation monitor. Your existing dbt project, scheduled next to SQL, notebooks, and AI agents with no separate runner.
The new Dashboard Builder in Databasin One turns a handful of plain-English questions into a coherent, multi-chart dashboard: modeled SQL, coordinated visuals, live preview. It then drops the result onto the drag-and-drop canvas for you to refine and publish.
Publish a dashboard to the Gallery with "Explore the data" enabled, and anyone viewing it gets a built-in chat locked to that dashboard's tables. The next question no longer needs a ticket to the data team.
Bring your XPS Shipper data into the lakehouse: bookings, tracking, and rate quotes, synced incrementally so every refresh only pulls what's new. Freight spend and carrier performance analytics, without the exports.
Group automation tasks into stages on a visual drag-and-drop board: everything in a stage runs concurrently, and each stage waits for the last to finish. Long sequential chains become fast, structured workflows.
Databasin's fourth SQL engine has arrived. Apache Doris brings low-latency OLAP analytics to the same open Iceberg tables you already query with Trino, Spark, and DuckDB: same editor, same streaming results, new speed profile.
Research institutions can now ingest the entire NIH RePORTER public dataset — no API keys, no setup friction. Millions of funded projects and publication linkages, refreshed on your schedule and ready for grants analytics.
Connect Databasin One directly to Power BI. Sign in, browse your workspaces, and pull semantic models — tables, relationships, and measures — straight into the AI's context. No more manual model exports.
Databasin's first multifamily property-management connector brings leasing, leases, households, operations, accounting, and maintenance data out of Fortress and into the lakehouse, with 14 curated views designed around how property teams actually work.
AI answers in Databasin One now cite their sources, and every citation is clickable: it opens a side panel showing the exact passage the answer came from, with the cited text highlighted, straight from your own ingested documents.
Finance teams can now sync Sage Intacct directly into Databasin. Eighteen accounting entities with hundreds of validated fields flow into analytics-ready tables, complete with a curated semantic model.
The Workday connector grows from 9 to 39 data feeds. Beyond core HCM, you can now ingest Financial Management, Resource Management, and Higher Education data — funds, grants, gifts, cost centers, projects, and academic appointees.
Databasin is now a full self-serve SaaS with radically transparent consumption pricing. Pay by the minute across five published rate categories — no seats, no commitments, no capacity pre-purchase. Stop using it, stop paying.
Databasin One's new persistent sidebar replaces the old setup wizard. Its headline feature, Discover, searches documents, files, and tables across your lakehouse and adds them to your AI's context in one click.
Build a production-grade CDC pipeline with Debezium 2.6, Kafka, and dbt 1.8 — idempotent MERGEs, backfill strategy, and monitoring that actually pages you.
Your Databasin One sessions no longer vanish when you close the tab. Workspaces persist your conversation, data context, charts, and generated documents. Share the whole analysis with your teammates.
Charts you create in conversation with Databasin One can now become real dashboards. Drag, resize, and arrange tiles on a canvas, export to PDF, and publish to your team's Gallery, where charts re-run live against your data.
Pipelines and automations now live together in a single Integrations page. See every data workflow in one place, filter by type, and create new ones from one button.
Describe who will use your data and what they need to answer. Databasin's AI drafts a full semantic model from your table schemas and renders it on a visual entity canvas you can edit. Business-ready models in minutes, not weeks.
The SQL editor no longer waits for your whole query to finish. Trino results now stream into the grid page by page, with live progress and stats — big queries feel instant.
The Databasin lakehouse is now truly multi-engine. Run Trino for federated SQL, Spark and Python for heavy processing and notebooks, or DuckDB for fast lightweight analytics — all over the same open Iceberg tables, switchable from a single selector.
Have an API we don't have a connector for yet? Drop in its OpenAPI/Swagger spec and Databasin builds the ingestion configuration for you: endpoints, pagination, and schemas included.
The 14-day clock is gone. New organizations now start with a credit balance and pay only for what they actually use — explore at your own pace, with full visibility into where every dollar goes.
A new command-line interface puts Databasin in your terminal — and pairs with Claude Code so AI coding agents can operate the platform on your behalf. Script pipeline operations, automate workflows, and let your AI assistant do the clicking.
Databasin One brings conversational AI to your lakehouse: ask questions in plain English, get live SQL-backed answers, charts, and exportable documents. It grows out of the AI Chat experience introduced in August and becomes the platform's AI home base.
Two heavyweight sources arrive in one month: pull operational data straight out of MongoDB, and bring HR data from SAP SuccessFactors into your lakehouse without custom code.
Databasin pipelines can now deliver data directly into Microsoft Fabric. Teams standardizing on Fabric get the same automated ELT experience already available for ADLS and Databricks targets.
Nothing matches that search. Try a broader term or clear the category filter.
Want to try any of this on your own data?