Skip to content
AskTable

Documentation feedback

Current pageYour first question/en/getting-started/quickstart

0 / 5,000

Your feedback will be sent to the maintenance team.

Go to console
Go to console

Using AskTable

Ask your first question

Ask a data question, inspect the results, and save them.

Guide · Module: getting-started · Users · Deployment: Not specified · Confirmed product versions: Not specified · Source: Existing documentation

Before you start

  • Your project must contain at least one data source whose metadata parsing is complete, and your role must have access to it.
  • No data yet? Ask the person responsible for data setup to connect a data source first.

Steps

Start a conversation

Click “New conversation” in the main navigation on the left. The center of the page contains the question area, with an input box and example questions below.

Select a data source

Click the data source selector below the input box (defaults to all data sources) and choose one from the list. The selector then shows your selection.

If no source is available, check the troubleshooting table below. If you manage data connections, you can also click “Create data source”.

What if the project uses data agents?

Use the agent selector instead. The agent determines which data sources to use, and the data source field indicates that the agent controls the selection.

Write your question

Specify the time range, grouping dimensions, filters, and sort order in the input box, then send.

For example, if your data contains stores, dates, and sales amounts, ask: “What were sales for each store last month? Sort from highest to lowest.” Replace the example fields with fields actually available in your data.

Wait for results

After sending, the message moves through thinking and analyzing states. You can see which tools are used, such as querying data, viewing data tables, and creating charts.

If more conditions are needed, the agent asks a clarifying question and lists options. Select your answers and click “Submit”, or click “Skip” if you do not want to answer.

Review and save results

Open “Table” in the results area and verify the time range, fields, and numbers. The table shows its current row and column counts and supports filtering, sorting, and row search. With chart visualization enabled, you can also view “Chart”.

To keep the result, use “Download CSV” to download a file or “Save to Tables” to create a reusable data table.

You are done when: you have viewed the result table for this question and retained it by downloading or saving. A chart is not required to complete your first question.

Key constraints

ItemValueWhere to change it
Run modeDeep thinking by default; enabling “Fast mode” gives faster but less detailed analysis“+” menu beside the input → “Fast mode”
Attachment uploadsExcel and CSV; defaults are Excel up to 10 MB per file, at most 20 sheets per file, and at most 200 fields (columns) per sheet; CSV up to 20 MB per file and at most 50 fields (columns); does not depend on the Python interpreter“+” menu to the left of the input
ChartsChart visualization must be enabled; otherwise results contain tables onlyAgent configuration → “Capabilities”
Clarifying questionsUser clarification must be enabledAgent configuration → “Capabilities”
Web searchOff by default; a system administrator must also configure it in system settings before it can be usedAgent configuration → “Capabilities”

Troubleshooting

Start with the symptoms below; look up other errors in Common questions.

SymptomHow to identify itWhat to do
No results for a long timeNo new tool calls for more than 1 minuteWait; if processing pauses or reaches the per-turn limit, click “Continue”
Cannot send a messageSending is rejected with a project-locked messageThe project's quota or license has been exhausted; contact an administrator about licensing
No data source to selectThe data source list is emptyThe project has no sources, or your role cannot access them; see Permissions and data scopes
Numbers differ from expectationsExpand the result table to see the SQL actually executedCheck the time range, granularity, and filters first; if the business definition is wrong, improve table and field descriptions
The agent asks a question backIt lists options to confirm instead of immediately returning resultsThe question is missing conditions. Select answers and click “Submit”, or include the definitions directly next time

Next steps

  • Revisit results regularly: save a data table and add it to a data dashboard
  • Investigate a complex question through attribution or drill-down: use an analysis canvas
  • Let your team ask questions in Feishu: see Feishu bot
  • Share results externally: use “Share conversation” → enable the public link switch