System Tasks

Available system tasks

System Task	Type	Description
HTTP	HTTP	Call any HTTP/REST endpoint. Supports GET, POST, PUT, DELETE with headers, body, and connection/read timeouts.
HTTP Poll	HTTP_POLL	Repeatedly call an HTTP endpoint until a specified condition is met — useful for polling a long-running external job.
gRPC	GRPC	Invoke a remote gRPC service endpoint from a workflow.
Inline	INLINE	Execute lightweight JavaScript or Python expressions server-side using GraalJS. Useful for data transformation, validation, and simple logic.
JSON JQ Transform	JSON_JQ_TRANSFORM	Transform JSON data using jq expressions. Powerful for reshaping, filtering, and aggregating data.
Business Rule	BUSINESS_RULE	Evaluate business rules defined in a spreadsheet-based decision table.
JDBC	JDBC	Execute SQL queries and updates against relational databases (MySQL, PostgreSQL, Oracle, etc.) with connection pooling and transaction management.
Event	EVENT	Publish events to external systems — Kafka, NATS, NATS Streaming, AMQP (RabbitMQ), SQS, or Conductor's internal queue.
Wait for Webhook	WAIT_FOR_WEBHOOK	Pause workflow execution until a matching webhook callback is received.
SendGrid	SENDGRID	Send transactional or notification emails through the SendGrid email service.
Update Task	UPDATE_TASK	Update the status of a different task in a running workflow.
Update Secret	UPDATE_SECRET	Update the value of an existing secret.
Get Signed JWT	GET_SIGNED_JWT	Generate a signed JSON Web Token (JWT) for authenticating with another system.
Opsgenie	OPS_GENIE	Create, close, or manage alerts in Opsgenie.
Query Processor	QUERY_PROCESSOR	Run queries against systems such as Conductor APIs and Prometheus metrics to evaluate conditions for alerting workflows.

AI & LLM tasks

Conductor provides native AI system tasks with direct integration with 14+ LLM providers and 4 vector databases — no external frameworks or custom workers needed.

Task	Type	Description
LLM Chat Complete	LLM_CHAT_COMPLETE	Multi-turn conversational AI with optional tool calling. Supports all major LLM providers.
LLM Text Complete	LLM_TEXT_COMPLETE	Single prompt completion.
LLM Generate Embeddings	LLM_GENERATE_EMBEDDINGS	Create vector embeddings from text or data using an AI model.
LLM Store Embeddings	LLM_STORE_EMBEDDINGS	Store generated embeddings in a vector database for later retrieval.
LLM Get Embeddings	LLM_GET_EMBEDDINGS	Retrieve stored embeddings from a vector database.
LLM Search Embeddings	LLM_SEARCH_EMBEDDINGS	Search a vector database using a precomputed embedding or a natural-language query.
LLM Index Text	LLM_INDEX_TEXT	Index text content into a vector database for semantic search.
LLM Index Document	LLM_INDEX_DOCUMENT	Index documents into a vector database for semantic search and retrieval.
LLM Search Index	LLM_SEARCH_INDEX	Query a vector database to find documents relevant to a given query.
Get Document	GET_DOCUMENT	Retrieve stored documents for further AI processing or analysis.
Parse Document	PARSE_DOCUMENT	Extract, convert, and structure documents into Markdown for AI processing.
Chunk Text	CHUNK_TEXT	Split large text into smaller segments for efficient AI processing.
List Files	LIST_FILES	Retrieve file listings from cloud storage, Git repositories, or websites for downstream processing.
Generate Image	GENERATE_IMAGE	Generate images from a text prompt using a configured AI model integration.
Generate Audio	GENERATE_AUDIO	Convert text into synthesized speech or audio using a configured AI model integration.
Generate Video	GENERATE_VIDEO	Generate video from a text prompt or an input image using a configured AI model integration.
Generate PDF	GENERATE_PDF	Convert Markdown content into a styled PDF document.
List MCP Tools	LIST_MCP_TOOLS	Connect to an MCP server and list the tools it exposes.
Call MCP Tool	CALL_MCP_TOOL	Invoke a specific tool exposed by an MCP server and return its result.

Supported providers: Anthropic (Claude), OpenAI (GPT), Azure OpenAI, Google Gemini, AWS Bedrock, Mistral, Cohere, HuggingFace, Ollama, Perplexity, Grok (xAI), and more. Switch providers by changing a configuration parameter — no code changes required.

Supported vector databases: Pinecone, Weaviate, pgvector (PostgreSQL), and MongoDB Atlas Vector Search. These enable RAG (retrieval-augmented generation) pipelines as standard Conductor workflows.