Why teams use it
Warehouses, Airflow, and docs often live in different tabs — onboarding slows down. Here you see structure and movement in one place: canvas, orchestration nearby, an assistant that uses what you actually connected.
One set of saved credentials. Same connections for schema, Flow, ETL, and the docked AI on data pages. Fewer exports from desktop tools that were old the day after.
Analysts read the map. Engineers check FK paths before SQL. Leads use one live picture in reviews.
Miro-style, but for databases and ETL
Database & pipeline canvas
Schema Explorer draws tables as nodes and foreign keys as edges. Save layout, mark roles, pull more of the model onto one board. The assistant reads your saved connections — not a generic chat.
Goal: from “new connection” to a first useful map in minutes — how bases connect, where ETL sits, where Airflow runs. Then you refine.
Illustrative layout — your tenant’s objects and edges come from live metadata.
Saved positions and roles help new teammates and preserve architectural intent as the model grows.
Open Schema ExplorerBuilt for mixed data teams
One surface for analytics engineers, software teams, and stakeholders who need a picture without SQL.
- Faster onboarding when the warehouse outpaced internal docs.
- Fewer “which table is source of truth?” loops before planning.
- One place to show auditors or partners a live metadata-backed story.
Workspace after sign-in
Workspace is the hub: AI balance, shortcuts to Airflow, Database Explorer, ETL, meetings, and your saved connections.
- Connections you add here match Database Explorer.
- Tiles keep data work and AI in one home page.
Flow — canvas with ETL overlay
Same database canvas, flow-oriented toolbar: ETL palette, edges into tables, saved layouts where we support them.
Bottom dock on data pages
On database, Flow, and ETL pages a strip at the bottom keeps tools in reach: switch connection, schema explorer, ETL, add connection, Ask AI — you stay on the canvas.
- ETL respects the selected connection from the dock.
- One strip for explorer, pipelines, and assistant on wide layouts.
Samples and previews
Table previews use bounded samples — enough to validate types without pulling full production into the browser. Filters stay compact for wide tables.
Connect engines in a few clicks
Add connection strings and credentials; separate entries per environment when that matches how you work. Supported families include PostgreSQL, ClickHouse, MySQL, SQL Server, and Airflow alongside databases.
Credentials stay under your platform account — you are not retyping secrets every session.
Jump from the canvas to the right engine without hunting spreadsheets for connection strings.
Your Airflow, readable next to the model
Point the app at your Airflow base URL (REST). You see DAG inventory, runs, tasks, and logs in the same mental space as your schemas — orchestration is not on a separate island.
Credentials follow the same patterns as other integrations (account-saved or managed defaults).
Apache Airflow
DAGs, schedules, task status — from koveh.com
Integration in short▼
The Airflow UI is read-only via a server proxy: your browser does not hold Airflow secrets. Airflow 2.x and 3-style APIs are supported depending on your deployment.
- DAG list, runs, task instances, logs, and DAG source where the API exposes it.
- Task graph from downstream metadata (with fallbacks when the list endpoint omits edges).
ETL and flows
ETL pages show how data moves — same account as the schema, no separate hunt for tools.
Jump from databases to pipelines without losing context.
ETL v2 (beta) is experimental; daily work stays on the main ETL page.
Assistant
AI with your data in context
On data pages the model uses your connection, schema, and small samples — answers stay tied to reality. Model choice, balance, longer sessions: open the AI workspace (button below).
- Browsing schemas and the canvas is free; AI usage is pay-as-you-go from the same balance you top up in Workspace.
- Docked modes on database pages keep Ask / Plan / Agent next to the canvas.
- Attachments and other flows have dedicated routes when you need files outside the dock.
From account to overview
Low friction: register, connect, explore.
- 01
Account
Create a user on koveh.com and sign in.
- 02
Connections
Add databases and optionally Airflow — separate entries per environment if needed.
- 03
Map & ask
Open Schema Explorer or Flow, arrange the view, then use Airflow, ETL, and AI from the same workspace.
Secrets and configuration
Keys belong in secure configuration (e.g. Infisical or runtime env), not committed source. Saved UI connections are tied to your account.
Data region & compliance
We design for GDPR-minded processing: customer data for the cloud product is hosted in Helsinki (Finland / EU). Koveh operates with Swiss structuring where applicable — we can walk through budget, invoicing, and compliance questions on a call. Nothing here is legal advice; we document what we implement and sign.
Azure-friendly deployment is on the roadmap — talk to us if you need a specific region or tenant model.
Two ways to work with Koveh
Same product: map data and pipelines, add AI when it helps. Price and hosting follow the path you pick.
koveh.com — cloud
Sign in, add connections, explore. Schemas, canvas, Flow, Airflow — free to browse. You pay only for AI (tokens), top-up in Workspace. We do not meter normal browsing. Heavy automation limits may come later; we will say so plainly.
Enterprise — on-prem or licensed package
For larger teams or regulated environments, we offer solution in-house: deploy in your perimeter (or dedicated tenancy), optional license-style delivery for our software, SSO and your policies, and professional services for integration. Contact us for architecture, security review, and rollout — we go deeper than this page.
Contact — enterprise & on-premYour infrastructure
Discuss VPC placement, secrets, backups, and support levels when you move from the public app to a controlled rollout.
Start a conversationThis page follows the site language toggle in the header.