Features
Qry is one workbench: what people ask, build and share runs under the same identity, permissions and models. These pages describe each part in detail.
Asking and sharing
- AI Analyst — ask in your own words; Qry writes and runs the SQL, charts the result and explains it, showing every step.
- Python execution — when SQL is not enough, Qry writes and runs Python in an isolated sandbox, with matplotlib and Plotly charts.
- File upload — attach CSV, Excel, PDF, images, Markdown or text to a conversation.
- Conversation checkpoints — rewind a conversation to an earlier point and take another path.
- Dashboards — built by asking, edited like a document: 17 chart types, KPIs with period comparison, drill-down, forecasts, anomaly detection and questions about what is on the page.
- Scheduled tasks — "every Monday at nine, send me sales by region", delivered to a conversation or by email.
- Messaging gateway — ask from Telegram, with the same permissions as the web.
Teams and knowledge
- Workspaces — a team's conversations, documents and memory; answers cite the passage they used.
- Skills — business rules, reference documents, procedures and scripts, applied by scope and versioned in Git.
- Domain context — the reference-document kind of Skill, in depth.
Building with data
- Notebooks — SQL, Python, PySpark and prompt cells, scheduled runs, an AI assistant and Git.
- Lakeflow — declarative pipelines and DAG jobs, including SQL that runs inside your warehouse.
- DataFlow — data movement between systems from plain-language instructions.
- Forge — migration of Oracle, Teradata, SQL Server and SAS estates, code included, with parity checks.
- Git folders — notebooks, pipelines, jobs and skills as files in GitHub, GitLab or Bitbucket.
- ML Hub — train, compare and monitor machine-learning models from a conversation.
- Nexus — curated data products and an API for other systems.
- Data profiling — what Qry learns of each table from a sample, so answers filter on values that exist.
Administration and security
- Security architecture — how every query is checked, and where the limits are.
- RBAC — grants from datasource down to table, for users and groups.
- ABAC — row filters and column rules by attribute.
- Database connectivity — the native connectors and how to set them up.
- Task quota templates — limits on scheduled work.
- MCP integration — Model Context Protocol servers as extra tools.
- DataHub integration — metadata from DataHub, filtered by permission.
- Issue reporting — where reported issues go.
- Claude context overflow — how long conversations stay within the context window.
Where next
- Getting started — a first question in a few minutes.
- User guide — step-by-step guides.
- Administrator guide — users, datasources, access control and models.