Nodes Catalog

Workflow Nodes

Complete catalog of all 58 custom nodes available in ConnectHealth workflows.

TL;DR / Key Takeaways

•58 pre-built nodes across AI Agent, FHIR, HL7, SMART Auth, Clearinghouse, and X12 categories.
•AI Agent nodes: LLM orchestration, tool execution, PHI guardrails, structured output, and conversation memory — powered by Bedrock, Azure OpenAI, or Vertex AI.
•Data transformation: FHIR Mapper, X12 Mapper, and HL7 Parser nodes convert between healthcare data standards without custom code.
•HTTP-in endpoints expose workflows as REST APIs — map any inbound request to FHIR queries, HL7 parsing, or AI processing.
•Clearinghouse and Bulk Export nodes handle real-time eligibility checks (EBVB) and large-scale FHIR data extraction.
•All nodes are drag-and-drop in the Workflow Builder with typed inputs/outputs for safe, validated connections.

AI Agent Nodes

Build intelligent agentic workflows with AI

AI Agent

An intelligent orchestrator node that processes messages using AI and can autonomously call tools to fulfill requests.

1 input→1 output

AI Model

Configuration node that defines AI model settings and credentials for use with AI Agent nodes. Supports multiple AI providers.

1 input→1 output

Guardrails

Applies intelligent safety guardrails to AI responses using Regex patterns or AI-enhanced logic for PHI redaction, content moderation, and quality checks.

1 input→2 outputs

AI Tool

Defines a custom tool that the AI Agent can call, and coordinates between the AI and the tool's execution flow.

1 input→2 outputs

AI Tool (Function)

Creates a JavaScript function-based tool that the AI Agent can call to execute custom logic.

1 input→1 output

AI HTTP Tool

Creates an HTTP request tool that the AI Agent can call to make external API calls with dynamic parameters.

1 input→2 outputs

Tool Terminate

Marks the end of a tool execution flow and returns the result back to the AI Agent.

1 input→0 outputs

AI Memory (File)

Stores AI conversation memories persistently in a JSON file. Memories survive Node-RED restarts.

1 input→1 output

AI Memory (In-Mem)

Stores AI conversation memories in volatile memory. Fast but data is lost on Node-RED restart.

1 input→1 output

FHIR & HL7 Nodes

Healthcare data integration nodes

Audit

Audit Log

Audit logging node for EHRConnect workflows. Captures a snapshot of the message payload, masks sensitive data, and writes a structured entry to AWS CloudWatch — without blocking the flow.

1 input→1 output

SMART Auth

Backend to Backend

Backend authentication node for system-level SMART on FHIR/EHR integrations.

1 input→1 output

config

Aws Config

Stores AWS connection credentials securely. This is the shared credential holder for AWS-backed operation nodes (SQS, DynamoDB, Athena) — reference it from a node's config dropdown instead of adding a per-service config node. Queue URLs, table names and other service settings stay on the operation node.

1 input→1 output

Dynamodb Config

Deprecated: use the shared AWS Config (ec-aws-config) for new DynamoDB nodes. Existing flows that reference this config keep working; to migrate, open each DynamoDB node and pick an AWS config.

1 input→1 output

Sqs Config

Deprecated: use the shared AWS Config (ec-aws-config) for new SQS nodes. Existing flows that reference this config keep working; to migrate, open each SQS node and pick an AWS config.

1 input→1 output

Twilio Config

Stores Twilio account credentials securely. Referenced by all other Twilio nodes.

1 input→1 output

AWS

AWS Athena Query

Runs a SQL query against an AWS Athena workgroup/database (tables registered in the Glue Data Catalog), replacing the discontinued-by-AWS S3 Select approach. Credentials come from the shared ec-aws-config node rather than a dedicated per-node config node.

1 input→2 outputs

DynamoDB Batch

Performs bulk DynamoDB operations — insert/delete many items at once (BatchWrite) or retrieve many items by key at once (BatchGet). Handles chunking and retry automatically.

1 input→2 outputs

DynamoDB Create

Inserts or replaces an item in a DynamoDB table. Used in ConnectHealth to store OAuth tokens, session data, FHIR resource cache, and audit logs.

1 input→2 outputs

DynamoDB Delete

Permanently removes an item by primary key. Used for session revocation, token expiry cleanup, and workflow termination.

1 input→2 outputs

DynamoDB Get

Retrieves a single item from a DynamoDB table by its exact primary key.

1 input→2 outputs

DynamoDB Query

Retrieves multiple items from a DynamoDB table using a key condition, with optional sort key filtering, attribute filtering, and pagination.

1 input→2 outputs

DynamoDB Scan

Reads every item in a DynamoDB table with optional attribute filtering. Appropriate for admin maintenance — finding expired tokens, inactive sessions, or orphaned records. Always prefer DynamoDB Query when you know the partition key.

1 input→2 outputs

DynamoDB Update

Updates specific attributes of an existing DynamoDB item without overwriting the entire record. Perfect for refreshing OAuth tokens or incrementing counters.

1 input→2 outputs

SQS Delete

Deletes one message, or many messages in a single batch call, from an SQS queue using their receipt handle(s). This permanently removes the message(s) from the queue — call it only after the message has been fully processed.

1 input→2 outputs

SQS Receive

Receives one or more messages from an AWS SQS queue, with optional long-polling and optional auto-delete on receipt.

1 input→2 outputs

SQS Send

Sends a single message to an SQS queue (standard or FIFO).

1 input→2 outputs

FHIR

Bulk Export Kickoff

Initiates a FHIR bulk data export for a specified group.

1 input→1 output

Bulk Export Retrieve

Downloads FHIR bulk export files and formats them for S3 upload.

1 input→1 output

Bulk Export Status

Checks the status of a FHIR bulk data export request.

1 input→1 output

FHIR Create

Creates a new FHIR resource on the server.

1 input→1 output

FHIR Delete

Deletes a specific FHIR resource by type and ID.

1 input→1 output

FHIR Read

Reads a specific FHIR resource by type and ID.

1 input→1 output

FHIR Search

Searches for FHIR resources by type and search parameters. Includes a built-in visual filter builder.

1 input→1 output

FHIR Search Filter

Builds search filter parameters for FHIR resources using a visual interface.

1 input→1 output

FHIR Update

Updates an existing FHIR resource on the server using a full resource replacement (PUT).

1 input→1 output

CCDA

CCDA Parse

Parses a CCDA (C-CDA) XML document — any genuine C-CDA document type (CCD, Discharge Summary, Referral Note, ...) — into structured JSON, with an optional FHIR Bundle output.

1 input→1 output

CDS Hooks

CDS Hooks In

Inbound trigger for a CDS Hooks invocation from the EHR — fires when the EHR calls this connection's invocation URL for the selected hook (patient-view, order-select or order-sign).

0 inputs→2 outputs

CDS Hooks Response

Completes a CDS Hooks flow — sends the response back to the EHR using the msg.ch.cds.responder callback attached by an upstream CDS Hooks In node.

1 input→1 output

sequence

Collect

A fan-in point. Every wire arriving at this node gets a slot; once every slot is filled, all of them are forwarded together as a single message.

1 input→1 output

Connection

Connection Parameter

Reads safe configuration values from a saved EHR connection and outputs them as an array on msg.payload.

1 input→1 output

HL7

HL7 Response

Sends an HL7 v2 ACK (acknowledgment) or NACK back to the hospital on the original MLLP socket. This is a terminal node — analogous to Node-RED's built-in http response node.

1 input→0 outputs

HL7 In

Receives inbound HL7 v2 messages from a hospital over a dedicated VPC connection (MLLP on port 2575).

0 inputs→1 output

HL7 Out

Sends an HL7 v2 message to a hospital's HL7 listener over a dedicated VPC connection using MLLP.

1 input→2 outputs

HL7 Parser

Parses HL7 messages into structured objects with metadata extraction.

1 input→1 output

Deprecated

FHIR Mapper (deprecated)

Deprecated. Use the Mapper node with a Mapping Project instead: it takes the raw message (HL7 v2, X12 or JSON) directly and needs no upstream parser. This node is kept only so existing flows keep working.

1 input→1 output

Mapping

Mapper

Runs a saved Mapping Project (authored in the visual Mapping editor). Accepts either a raw message or data an upstream node has already parsed. On Auto it selects the project's matching mapping by trigger event + version; you can instead pin one specific mapping.

1 input→1 output

MCP

MCP Response

Marks the end of an MCP tool flow and returns the result back to the AI client.

1 input→0 outputs

MCP Tool

Defines a tool that AI clients (Claude Desktop, Claude Code, Cursor) can discover and call over MCP, and starts this flow each time the AI calls it.

0 inputs→1 output

EMR API

EMR Connector

Calls a vendor's proprietary (non-FHIR) REST API — the counterpart to the FHIR nodes. Vendor-agnostic: it resolves the bound connection, selects the proprietary service for that EHR, and invokes a catalogued endpoint by id. First supported vendor: Open Dental.

1 input→1 output