Test Tool Connection

Test a tool connection.

Parameters:

For ToolConfigPostgres:

  • type: Literal type indication for Postgres tool (Postgres).
  • connection_string: Connection string for the Postgres database.

For ToolConfigMySQL:

  • type: Literal type indication for MySQL tool (MySQL).
  • connection_string: Connection string for the MySQL database.

For ToolConfigMSSQL:

  • type: Literal type indication for MSSQL tool (MSSQL).
  • connection_string: Connection string for the MSSQL database.

For ToolConfigOracle:

  • type: Literal type indication for Oracle tool (Oracle).
  • connection_string: Connection string for the Oracle database.

For ToolConfigBigQuery:

  • type: Literal type indication for BigQuery tool (BigQuery).
  • project_id: Project ID for BigQuery.
  • dataset_id: Dataset ID for BigQuery.
  • service_account: Service account information as a dictionary with keys and values as strings.

For ToolConfigPGVector:

  • type: Literal type indication for PG Vector tool (PGVector).
  • connection_string: Connection string for the PG Vector database.
  • table_name: Name of the table in the PG Vector database.
  • embedding_model: Configuration of the embedding model to be used.
    • provider: Provider of the embedding model (VertexAIModelGarden).
    • project_id: Project ID for Vertex AI.
    • endpoint_id: Endpoint ID for the model.
    • location: Location of the Vertex AI endpoint.
    • service_account: Service account information as a dictionary with keys and values as strings.
Recent Requests
Log in to see full request history
TimeStatusUser Agent
Retrieving recent requests…
LoadingLoading…

Tests the connection for a tool an agent can use. Supported tools are SQL databases (Postgres, MySQL, MSSQL, and Oracle), BigQuery, vector stores (PGVector and Pinecone), OpenAPI tools, MCP servers, and knowledge.

When to use it

Use it to validate connection details and credentials before you add the tool to an agent version.

Required parameters

  • Request body: the tool configuration. Its type selects the tool, and the remaining fields are that tool's settings, such as a connection_string for a SQL database.

Response

A successful call returns 204 No Content with an empty body.

Related: Test a model connection, Test MCP connectivity, Create an agent version

Body Params
const
enum
Defaults to PGVector
Allowed:
string
required
string
required
embedding_model
object
required
Responses
204

Successful Response

422

Unprocessable Entity

500

Internal Server Error

Language
Credentials
OAuth2
LoadingLoading…
Response
Click Try It! to start a request and see the response here!