Skip to main content

Module airbyte_agent_sdk.connectors.exa.connector

Exa connector.

Classes​

ContentsQuery(connector: ExaConnector) : Query class for Contents entity operations.

Initialize query with connector reference.

Methods​

list(self, urls: list[str], text: Any | None = None, highlights: Any | None = None, summary: ContentsListParamsSummary | None = None, **kwargs) ‑> airbyte_agent_sdk.connectors.exa.models.ExaExecuteResult[list[SearchResult]] : Get the full page contents, summaries, and metadata for a list of URLs. Returns instant results from Exa's cache, with automatic live crawling as fallback for uncached pages. Use this to retrieve text, highlights, and summaries for specific URLs.

Args: urls: Array of URLs to retrieve contents for. text: Text extraction options. Pass true for defaults or an object for advanced options. highlights: Highlight extraction options. Pass true for defaults or an object for advanced options. summary: Summary generation options. **kwargs: Additional parameters

Returns: ContentsListResult

ExaConnector(auth_config: ExaAuthConfig | AirbyteAuthConfig | BaseModel | None = None, on_token_refresh: Any | None = None) : Type-safe Exa API connector.

Auto-generated from OpenAPI specification with full type safety.

Initialize a new exa connector instance.

Supports both local and hosted execution modes:

  • Local mode: Provide connector-specific auth config (e.g., ExaAuthConfig)
  • Hosted mode: Provide AirbyteAuthConfig with client credentials and either connector_id or workspace_name

Args: auth_config: Either connector-specific auth config for local mode, or AirbyteAuthConfig for hosted mode on_token_refresh: Optional callback for OAuth2 token refresh persistence. Called with new_tokens dict when tokens are refreshed. Can be sync or async. Example: lambda tokens: save_to_database(tokens) Examples:

Local mode (direct API calls)

connector = ExaConnector(auth_config=ExaAuthConfig(api_key="..."))

Hosted mode with explicit connector_id (no lookup needed)

connector = ExaConnector( auth_config=AirbyteAuthConfig( airbyte_client_id="client_abc123", airbyte_client_secret="secret_xyz789", connector_id="existing-source-uuid" ) )

Hosted mode with lookup by workspace_name

connector = ExaConnector( auth_config=AirbyteAuthConfig( workspace_name="user-123", organization_id="00000000-0000-0000-0000-000000000123", airbyte_client_id="client_abc123", airbyte_client_secret="secret_xyz789" ) )

Class variables​

connector_name : The type of the None singleton.

connector_version : The type of the None singleton.

sdk_version : The type of the None singleton.

Static methods​

tool_utils(func: _F | None = None, *, update_docstring: bool = True, max_output_chars: int | None = 100000, framework: FrameworkName | None = None, internal_retries: int = 0, should_internal_retry: Callable[[Exception, tuple[Any, ...], dict[str, Any]], bool] | None = None, exhausted_runtime_failure_message: Callable[[Exception, tuple[Any, ...], dict[str, Any]], str | None] | None = None) ‑> ~_F | Callable[[~_F], ~_F] : Decorator that adds tool utilities like docstring augmentation and output limits.

Composes :func:airbyte_agent_sdk.translation.translate_exceptions for runtime wrapping (sync/async branch + output-size check + framework signal translation + optional internal retry loop), and adds connector-specific docstring augmentation on top of it.

Usage: @mcp.tool() @ExaConnector.tool_utils async def execute(entity: str, action: str, params: dict): ...

@mcp.tool() @ExaConnector.tool_utils(update_docstring=False, max_output_chars=None) async def execute(entity: str, action: str, params: dict): ...

@mcp.tool() @ExaConnector.tool_utils(framework="pydantic_ai", internal_retries=2) async def execute(entity: str, action: str, params: dict): ...

Args: update_docstring: When True, append connector capabilities to doc. max_output_chars: Max serialized output size before raising. Use None to disable. framework: One of "pydantic_ai" | "langchain" | "openai_agents" | "mcp". Defaults to None → auto-detect by attempting each framework's canonical import in order. Explicit always wins. internal_retries: How many transient runtime failures (429/5xx, network, timeout) to retry silently before surfacing. Default 0. Forwarded to :func:airbyte_agent_sdk.translation.translate_exceptions. should_internal_retry: Optional predicate (error, args, kwargs) -> bool further restricting which retryable errors are safe for this specific tool. Forwarded to :func:airbyte_agent_sdk.translation.translate_exceptions. exhausted_runtime_failure_message: Optional callback (error, args, kwargs) -> str | None. Invoked after internal retries are exhausted OR were skipped via should_internal_retry returning False. Forwarded to :func:airbyte_agent_sdk.translation.translate_exceptions.

Instance variables​

connector_id: str | None : Get the connector/source ID (only available in hosted mode).

Returns: The connector ID if in hosted mode, None if in local mode.

Methods​

check(self) ‑> airbyte_agent_sdk.connectors.exa.models.ExaCheckResult : Perform a health check to verify connectivity and credentials.

Executes a lightweight list operation (limit=1) to validate that the connector can communicate with the API and credentials are valid.

Returns: ExaCheckResult with status ("healthy" or "unhealthy") and optional error message

Example: result = await connector.check() if result.status == "healthy": print("Connection verified!") else: print(f"Check failed: {result.error}")

close(self) : Close the connector and release resources.

entity_schema(self, entity: str) ‑> dict[str, typing.Any] | None : Get the JSON schema for an entity.

Args: entity: Entity name (e.g., "contacts", "companies")

Returns: JSON schema dict describing the entity structure, or None if not found.

Example: schema = connector.entity_schema("contacts") if schema: print(f"Contact properties: {list(schema.get('properties', {}).keys())}")

execute(self, entity: str, action: "Literal['list']", params: Mapping[str, Any] | None = None, *, select_fields: list[str] | None = None, exclude_fields: list[str] | None = None, skip_truncation: bool = True) ‑> Any : Execute an entity operation with full type safety.

This is the recommended interface for blessed connectors as it:

  • Uses the same signature as non-blessed connectors
  • Provides full IDE autocomplete for entity/action/params
  • Makes migration from generic to blessed connectors seamless

Args: entity: Entity name (e.g., "customers") action: Operation action (e.g., "create", "get", "list") params: Operation parameters (typed based on entity+action) select_fields: Optional allowlist of dot-notation fields to include exclude_fields: Optional blocklist of dot-notation fields to remove skip_truncation: Disable long-text truncation for collection actions

Returns: Typed response based on the operation

Example: customer = await connector.execute( entity="customers", action="get", params={"id": "cus_123"} )

list_entities(self) ‑> list[dict[str, typing.Any]] : Get structured data about available entities, actions, and parameters.

Returns a list of entity descriptions with:

  • entity_name: Name of the entity (e.g., "contacts", "deals")
  • description: Entity description from the first endpoint
  • available_actions: List of actions (e.g., ["list", "get", "create"])
  • parameters: Dict mapping action -> list of parameter dicts

Example: entities = connector.list_entities() for entity in entities: print(f"{entity['entity_name']}: {entity['available_actions']}")

SearchResultsQuery(connector: ExaConnector) : Query class for SearchResults entity operations.

Initialize query with connector reference.

Methods​

list(self, query: str, type: str | None = None, category: str | None = None, num_results: int | None = None, include_domains: list[str] | None = None, exclude_domains: list[str] | None = None, start_published_date: str | None = None, end_published_date: str | None = None, start_crawl_date: str | None = None, end_crawl_date: str | None = None, contents: SearchResultsListParamsContents | None = None, moderation: bool | None = None, **kwargs) ‑> airbyte_agent_sdk.connectors.exa.models.ExaExecuteResult[list[SearchResult]] : Perform a search with an Exa prompt-engineered query and retrieve a list of relevant results. Optionally request contents (text, highlights, summary) inline with the search results. Supports filtering by domain, date, category, and number of results.

Args: query: The search query string. type: The type of search. auto intelligently selects the best mode, instant provides lowest latency, fast uses lower-latency models, deep-lite provides lightweight synthesis, deep performs in-depth research with synthesis, and deep-reasoning adds more reasoning for complex searches. category: A data category to focus on for improved result quality. num_results: Number of results to return (max 100). include_domains: List of domains to include. If specified, results will only come from these domains. exclude_domains: List of domains to exclude. If specified, no results will be returned from these domains. start_published_date: Only return links published after this date. ISO 8601 format. end_published_date: Only return links published before this date. ISO 8601 format. start_crawl_date: Only return links crawled by Exa after this date. ISO 8601 format. end_crawl_date: Only return links crawled by Exa before this date. ISO 8601 format. contents: Options for requesting page contents inline with search results. moderation: Enable content moderation to filter unsafe content. **kwargs: Additional parameters

Returns: SearchResultsListResult

SimilarResultsQuery(connector: ExaConnector) : Query class for SimilarResults entity operations.

Initialize query with connector reference.

Methods​

list(self, url: str, num_results: int | None = None, include_domains: list[str] | None = None, exclude_domains: list[str] | None = None, start_published_date: str | None = None, end_published_date: str | None = None, start_crawl_date: str | None = None, end_crawl_date: str | None = None, contents: SimilarResultsListParamsContents | None = None, **kwargs) ‑> airbyte_agent_sdk.connectors.exa.models.ExaExecuteResult[list[SearchResult]] : Find web pages similar to a given URL. Uses Exa's embeddings to find semantically similar content. Supports filtering by domains and dates.

Args: url: The URL to find similar pages for. num_results: Number of similar results to return (max 100). include_domains: List of domains to include. If specified, results will only come from these domains. exclude_domains: List of domains to exclude. If specified, no results will be returned from these domains. start_published_date: Only return links published after this date. ISO 8601 format. end_published_date: Only return links published before this date. ISO 8601 format. start_crawl_date: Only return links crawled by Exa after this date. ISO 8601 format. end_crawl_date: Only return links crawled by Exa before this date. ISO 8601 format. contents: Options for requesting page contents inline with similar page results. **kwargs: Additional parameters

Returns: SimilarResultsListResult

Was this page helpful?