Component Reference

Component Reference

RocketRide Engine includes 50+ pipeline nodes organized by category. Each node is implemented in Python and configured via a services.json file. See Pipeline API for how to use these components in pipeline definitions.

Node source code is located in nodes/src/nodes/<node_name>/.

LLMs

Large Language Models for text generation, summarization, question answering, and chat.

Node	Title	Protocol	Input Lanes	Output Lanes
llm_openai	LLM - OpenAI	llm_openai://	questions	answers
llm_anthropic	LLM - Anthropic	llm_anthropic://	questions	answers
llm_gemini	LLM - Gemini	llm_gemini://	questions	answers
llm_mistral	LLM - Mistral AI	llm_mistral://	questions	answers
llm_bedrock	LLM - Amazon Bedrock	llm_bedrock://	questions	answers
llm_ollama	LLM - Ollama	llm_ollama://	questions	answers
llm_deepseek	LLM - Deepseek	llm_deepseek://	questions	answers
llm_xai	LLM - xAI	llm_xai://	questions	answers
llm_perplexity	LLM - Perplexity	llm_perplexity://	questions	answers
llm_ibm_watson	LLM - IBM Watson	(internal)	questions	answers
llm_vertex	LLM - Vertex AI	(internal)	questions	answers

Vector Stores

Vector databases for storing embeddings and performing similarity search.

Node	Title	Protocol	Input Lanes	Output Lanes
chroma	Vector Store - Chroma	chroma://	documents, questions	documents, answers, questions
pinecone	Vector Store - Pinecone	pinecone://	documents, questions	documents, answers, questions
qdrant	Vector Store - Qdrant	qdrant://	documents, questions	documents, answers, questions
weaviate	Vector Store - Weaviate	weaviate://	documents, questions	documents, answers, questions
milvus	Vector Store - Milvus	milvus://	documents, questions	documents, answers, questions
astra_db	Vector Store - Astra DB	astra_db://	documents, questions	documents, answers, questions
atlas	Vector Store - MongoDB Atlas	mongodb+srv://	documents, questions	documents, answers, questions
vectordb_postgres	Vector Store - PostgreSQL	postgres://	documents, questions	documents, answers, questions

Embeddings

Convert text or images into vector representations for similarity search.

Node	Title	Protocol	Input Lanes	Output Lanes
embedding_openai	Embedding - OpenAI	embedding_openai://	documents, questions	documents, questions
embedding_transformer	Embedding - Transformer	embedding_transformer://	documents, questions	documents, questions
embedding_image	Embedding - Image	embedding_image://	documents	documents

Preprocessors

Split and chunk text for downstream embedding and storage.

Node	Title	Protocol	Input Lanes	Output Lanes
preprocessor_langchain	Preprocessor - General Text	preprocessor_langchain://	text, table	documents
preprocessor_code	Preprocessor - Code	preprocessor_code://	text	documents
preprocessor_llm	Preprocessor - LLM	preprocessor_llm://	text	documents

The General Text preprocessor supports multiple splitter strategies: Recursive Character, Character, Markdown, LaTeX, NLTK, spaCy, and custom.

Document Processing

Parse, extract, and transform documents.

Node	Title	Protocol	Description
autopipe	Parse/Process/Embed	autopipe://	All-in-one: parse, preprocess, and embed in a single node
llamaparse	Data - LlamaParse	llamaparse://	Advanced document parsing via the LlamaParse service
reducto	Data - Reducto	reducto://	Document reduction and extraction
ocr	Image - OCR	ocr://	Optical character recognition for images
extract_data	Text - Data Extractor	extract_data://	Structured data extraction from text

Text Analysis

Analyze, transform, and enrich text.

Node	Title	Protocol	Description
ner	NER - Named Entity Recognition	ner://	Extract named entities (people, places, organizations)
anonymize	Text - Anonymize	anonymize_text://	Remove or mask PII from text
summarization	Text - Summarization: LLM	summarization://	LLM-powered text summarization
dictionary	Text - Dictionary	dictionary://	Dictionary-based text lookup and enrichment
prompt	Text - Prompt	prompt://	Prompt template management
question	Text - Question	question://	Question construction and formatting
vectorizer	Vectorizer	vectorizer://	Text vectorization utilities

Media

Process images, audio, and video.

Node	Title	Protocol	Description
audio_transcribe	Audio - Transcribe	audio_transcribe://	Transcribe audio to text
audio_player	Audio - Player	audio_player://	Audio playback control
frame_grabber	Video - Frame Grabber	frame_grabber://	Extract frames from video
image_cleanup	Image - Cleanup	image_cleanup://	Image preprocessing and cleanup
thumbnail	Image - Thumbnail	thumbnail://	Generate image thumbnails
llm_vision_mistral	Image - Mistral Vision	image_vision_mistral://	Mistral-based image understanding

Output

Return results to clients or write to storage.

Node	Title	Protocol	Description
response	HTTP Results	response://	Return structured results to the calling client
text_output	Text Output	text-output://	Format text output for client responses
local_text_output	Local Text Output	local-text-output://	Write text results to the local filesystem

Infrastructure & Storage

Database and storage connectors.

Node	Title	Protocol	Description
db_mysql	Database - MySQL	db_mysql://	MySQL database integration
index_search	Vector Store - Elasticsearch	elasticsearch://	Vector storage and retrieval with Elasticsearch (Cloud Serverless, Cloud Hosted, self-managed). Supports both BM25 text search and vector similarity search.
index_search	OpenSearch	opensearch://	BM25 and vector search for self-managed OpenSearch. Supports index mode and vector store mode.
remote	Remote	(internal)	Remote node execution
webhook	Webhook	(internal)	HTTP webhook source

Utility / Internal

Shared base classes and utilities used by other nodes.

Node	Description
llm_base	Base class for all LLM nodes
core	Shared core utilities
library	Shared library functions

Node File Structure

Each node directory contains:

nodes/src/nodes/<node_name>/
├── services.json       # Metadata, config schema, lanes, profiles, test cases
├── requirements.txt    # Python dependencies
├── __init__.py         # Python package init
├── <node_name>.py      # Main implementation
├── README.md           # Optional documentation
└── tests/              # Optional unit tests

Next Steps

• Pipeline API -- how to reference these components in pipelines
• Data Flow -- how data moves between components
• Module Guide -- build and test commands