Skip to main content
POST
/
v1
/
tools
/
database-query
Database Query
curl --request POST \
  --url https://{base_url}/gpt/v1/tools/database-query \
  --header 'Content-Type: application/json' \
  --data '
{
  "query": "<string>",
  "artifacts": [
    {
      "value": "<string>",
      "type": "file"
    }
  ]
}
'
{
  "content": [
    {
      "type": "text",
      "text": "The total revenue for Q4 was $3.5M, with the highest contributions from the Sales department."
    }
  ],
  "is_error": false
}

Body

application/json

Request body for querying connected SQL databases using natural language. Requires at least one SQL database artifact in the tool context.

Request body for database query using the tool.

​
query
string
required

The natural language query to run against connected databases.

Example:

"What were the Q4 revenue trends?"

​
artifacts
(FileArtifact · object | UriArtifact · object | TextArtifact · object | IngestedArtifact · object | SqlDatabaseArtifact · object)[]
required

List of SQL database artifacts to query against. At least one artifact of type 'sql_database' is required.

Input for base64 encoded files.

Response

Database query completed successfully.

Response returned from tool-based operations.

​
content
(TextBlock · object | ImageBlock · object | AudioBlock · object | BinaryBlock · object | ResourceLinkBlock · object | ResourceBlock · object | SourceBlock · object | ToolUseBlock · object | ThinkingBlock · object | TLDRBlock · object)[]
required

List of content blocks generated by the tool. Blocks can include plain text, citations, source attributions, or images.

Text content block containing natural language text.

​
is_error
boolean
default:false

True if the tool encountered an error during execution.