01
Connect your data
Attach a supported database to an Agent and choose which tables it can use.
- PostgreSQL and MySQL support
- Read-only database credentials recommended
- Test connections before use
How it works
From a business question to a grounded data answer
DataAgent combines your database schema, business knowledge, and an AI analysis engine to plan, query, compute, and explain—without exposing the underlying model provider.
Your question
DataAgent Analysis Engine
Schema + business context
Tables, charts, explanation
The workflow
A clear path from connection to decision
02
Teach DataAgent your business
Add context that raw column names cannot provide.
- Metric and field definitions
- Business rules, FAQs, and documents
- Logical relationships and semantic models
03
Ask in plain language
Describe the business question instead of writing SQL or analysis code.
- Preset questions for repeat workflows
- Follow-up questions within a session
- Optional human review mode
04
Plan, query, and analyze
DataAgent selects the required steps and uses live query results as evidence.
- SQL generation and execution
- Python analysis when computation is needed
- Retry and correction inside the workflow
05
Receive a usable answer
Results are organized for people who need to understand and act on them.
- Result tables and charts
- HTML or Markdown reports
- Explanations grounded in returned data
Inside the product
What DataAgent helps your team do
The interface is organized around the work—not around the underlying AI technology.
Query live databases
Work from permitted tables in connected business databases.
Generate SQL and analysis
Turn a question into query and computation steps without manual coding.
Use business knowledge
Apply definitions, rules, and documents when interpreting results.
Build visual outputs
Present result sets, charts, and structured explanations.
Keep a review checkpoint
Use human approval before execution when your workflow requires it.
Choose your AI mode
Use DataAgent built-in AI or connect your own compatible chat model.
Control and boundaries
Designed for understandable operation
DataAgent separates the AI provider from knowledge retrieval and keeps high-risk model internals out of day-to-day setup.
- Built-in knowledge search is managed automatically
- Only one chat provider is active per tenant
- BYOK connections are tested before activation
- Database access follows the credentials you provide
AI assistants
Bring your own AI assistant
DataAgent speaks MCP — the open standard for AI tools. Connect Claude Desktop or Cursor to an analyst that knows your business: your documents, your definitions, your metrics.
Analysts customized per business
Agent Knowledge, Business Knowledge and Semantic Models shape every answer with your context — not generic SQL over raw tables.
Full analyses, not just queries
Ask your AI assistant a question and get a finished report: SQL, execution, charts and narrative — inside the same validated pipeline the web app uses.
Scoped, metered, revocable
Access tokens are read-only by default, tenant-isolated on every call, usage-counted and revocable in one click.
Ready to ask your data a real question?
Start free, connect a database, and build your first Agent.