pipeline.add(name="...").<node>(...). Each entry lists the node’s configuration parameters. See the Pipeline reference for add, run, and lifecycle methods.
chunking — Chunking
Split text into chunks. Supports different chunking strategies like markdown-aware, sentence-based, or dynamic sizing.
str
default:"'markdown'"
Strategy for grouping segmented text into final chunks. ‘sentence’: groups sentences; ‘markdown’: respects Markdown structure (headers, code); ‘dynamic’: optimizes breaks for size using chosen segmentation method.
One of:
dynamic, markdown, sentencestr
default:"''"
The text to chunk
int
default:"0"
The overlap of each chunk of text.
int
default:"512"
The size of each chunk of text.
str
default:"'words'"
The method to break text into units before chunking. ‘words’: splits by word; ‘sentences’: splits by sentence boundary; ‘paragraphs’: splits by blank line/paragraph.
One of:
paragraphs, sentences, wordscreate_workspace — Create Workspace
Create a new workspace in a portal, upload files to its knowledge base, and share with users
Platform docs: Create Workspace
AcceptsFileList
default:"[]"
str
required
AcceptsPortal
required
knowledge_base — Knowledge Base
Semantically query a knowledge base that can contain files, scraped URLs, and data from synced integrations (e.g., Google Drive).
Platform docs: Knowledge Base
bool
default:"False"
Use additional LLM calls to analyze each document to improve answer correctness
bool
default:"False"
Filter the content returned from the knowledge base. Agents should provide structured metadata filters directly in the filter input when useful.
bool
default:"False"
Enable context
bool
default:"True"
Format the context for the LLM
bool
default:"False"
Enable the document DB filter
bool
default:"False"
Generate an LLM response from the retrieved context
bool
default:"False"
Whether to stream the LLM response
str
required
The query will be used to search documents for relevant content semantically. Must not be empty, only include relevant information for retrieval or metadata filter generation. Generally expand any specific acronyms or abbreviations but include the original acronym or abbreviation as well
str
default:"''"
Additional context to pass to the query analysis and qa steps
str
default:"''"
Filter the documents returned from the knowledge base
str
default:"''"
Structured metadata filter JSON for the knowledge base query. Use a top-level boolean clause such as {“type”:“condition”,“field”:“title”,“operator”:“match”,“value”:“Q4 report”}; leave empty when no hard metadata constraint is needed.
bool
default:"False"
Use an LLM to generate metadata filters to refine your query. Agents should usually leave this false and provide filters directly in the filter input.
AcceptsKnowledgeBase
required
Select an existing knowledge base, Use $object.knowledge_base.? syntax
str
default:"''"
The system prompt to use for the LLM
RetrievalConfig
RerankConfig
QueryEnhancementConfig
float
The alpha value for the retrieval. 1.0 is pure vector search and 0.0 is pure lexical search
bool
Extract separate questions from the query and retrieve content separately for each question to improve search performance
bool
Do a natural language metadata query
bool
Expand query to improve semantic search
bool
Expand query terms to improve semantic search
int
The number of chunks to rerank
bool
Rerank the documents returned from the knowledge base
str
Refine the initial ranking of returned chunks based on relevancy
str
The unit of retrieval. Chunks will return the most relevant chunks from the knowledge base as well as their text content. Documents will return the document metadata as well as most relevant snippets from the document. Pages will return complete pages with all chunks from pages containing relevant content
One of:
chunks, documents, pagesfloat
The score cutoff
int
The number of relevant chunks to be returned
bool
Transform the query for better semantic search
str
default:"'accurate'"
The mode to use for the advanced search
One of:
accurate, faststr
default:"'gpt-4o-mini'"
The model to use for the QA
knowledge_base_actions — Knowledge Base Actions
Create, load, and sync Knowledge Bases
str
default:"''"
knowledge_base_agent — Knowledge Base Agent
Query a knowledge base using an agentic approach with tools.
Platform docs: Knowledge Base Agent
str
default:"'google'"
Select the LLM provider to be used by the agent
One of:
googlestr
default:"'focused'"
Controls the query effort: ‘fast’ for quick answers, ‘focused’ for balanced depth, ‘deep’ for thorough analysis
One of:
deep, fast, focusedbool
default:"False"
If enabled, shows an additional context input to provide context to the agent
bool
default:"False"
If enabled, returns the relevant context/chunks used to generate the answer
str
default:"'gemini-3-flash-preview'"
Select the LLM model to be used by the agent
One of:
gemini-3-flash-previewstr
required
The natural language query. The agent will use this to determine the best way to query the knowledge base. Include the key criteria needed to answer the query.
str
default:"''"
Optional additional context to help the agent understand the query better (e.g., conversation history, user preferences)
AcceptsKnowledgeBase
required
Select an existing knowledge base. You must provide the id in $.object.knowledge_base.id format
bool
default:"True"
If enabled, generates a synthesized answer from the knowledge base
knowledge_base_create — Knowledge Base Create
Dynamically create a Knowledge Base with configured options
Platform docs: Knowledge Base Create
str
default:"'advanced'"
Strategy for grouping segmented text into final chunks. ‘sentence’: groups sentences; ‘markdown’: respects Markdown structure (headers, code); ‘dynamic’: optimizes breaks for size using chosen segmentation method.
One of:
advancedbool
required
To analyze document contents and enrich them when parsing
str
default:"''"
Apify API Key for scraping URLs (optional)
int
required
The overlap of the chunks to store in the knowledge base
int
required
The size of the chunks to store in the knowledge base
str
required
The name of the collection to store the knowledge base in
str
required
The embedding model to use for the knowledge base. Format: provider/model
str
required
The embedding provider to use
str
required
The file processing implementation to use for parsing documents
One of:
contextual_ai, default, docling, llama_parse, mistral_ocr, reducto, textractbool
required
Whether to create a hybrid knowledge base
str
required
The name of the knowledge base to create
str
required
The precision to use for the knowledge base
str
default:"'words'"
The method to break text into units before chunking. ‘words’: splits by word; ‘sentences’: splits by sentence boundary; ‘paragraphs’: splits by blank line/paragraph.
One of:
paragraphs, sentences, wordsbool
required
Whether to shard the knowledge base
str
required
The vector database provider to use
knowledge_base_fetch_document_content — Knowledge Base Fetch Document Content
Fetch the full content of a specific document from a knowledge base by scrolling through all its chunks
Platform docs: Knowledge Base Fetch Document Content
str
required
AcceptsKnowledgeBase
required
int
default:"200"
int
default:"0"
str
default:"''"
knowledge_base_fetch_items — Knowledge Base Fetch Items
Advanced knowledge base item fetching with traversal, filtering, and output shaping capabilities
Platform docs: Knowledge Base Fetch Items
str
default:"''"
str
default:"'ALL'"
One of:
ALL, DOCUMENTS, FOLDERSAcceptsKnowledgeBase
required
int
default:"50"
int
default:"0"
int
default:"0"
str
default:"''"
str
default:"''"
str
default:"''"
str
default:"'metadata'"
One of:
long, metadata, shortknowledge_base_get_item_bboxes — Knowledge Base Get Item Bboxes
Fetch OCR bounding boxes for specific pages of a PDF document in a knowledge base
str
required
AcceptsKnowledgeBase
required
str
required
knowledge_base_list_items — Knowledge Base List Items
List items (documents and folders) from a knowledge base with pagination support
AcceptsKnowledgeBase
required
int
default:"50"
bool
default:"False"
int
default:"0"
knowledge_base_loader — Knowledge Base Loader
Load data into an existing knowledge base.
Platform docs: Knowledge Base Loader
str
default:"'File'"
Select the type of data to load
One of:
File, URLbool
default:"False"
Scrape sub-pages of the provided link
AcceptsKnowledgeBase
required
The knowledge base to load data into
str
default:"'Never'"
The frequency to rescrape the URL
One of:
Daily, Monthly, Never, Weeklystr
required
The raw URL link (e.g., https://vectorshift.ai/)
bool
default:"False"
Use a proxy to crawl the website
bool
default:"False"
Load URLs to crawl from a sitemap. If the URL is a sitemap, it will be used directly. If the URL is not a sitemap, the sitemap will be fetched automatically.
int
default:"5"
The maximum depth of the URL to crawl
int
default:"10"
The maximum number of recursive URLs to scrape
bool
default:"False"
Whether to only crawl links from the same domain
AcceptsFileList
required
The file to be added to the selected knowledge base. Note: to convert text to file, use the Text to File node
knowledge_base_sync — Knowledge Base Sync
Automatically trigger a sync to the integrations in the selected knowledge base
Platform docs: Knowledge Base Sync
AcceptsKnowledgeBase
required
semantic_search — Semantic Search
Generate a temporary vector database at run-time and retrieve the most relevant pieces from the documents based on the query.
Platform docs: Semantic Search
bool
default:"False"
Use additional LLM calls to analyze each document to improve answer correctness
bool
default:"False"
Filter the content returned from the knowledge base
bool
default:"False"
Additional context passed to advanced search and query analysis
bool
default:"False"
Format the context for the LLM
bool
default:"False"
Filter the documents returned from the knowledge base
str
default:"'markdown'"
Strategy for grouping segmented text into final chunks. ‘sentence’: groups sentences; ‘markdown’: respects Markdown structure (headers, code); ‘dynamic’: optimizes breaks for size using chosen segmentation method.
One of:
dynamic, markdown, sentencestr
default:"'openai/text-embedding-3-small'"
The model to use for the embedding
str
required
The query will be used to search documents for relevant pieces semantically.
bool
default:"False"
To analyze document contents and enrich them when parsing
str
default:"''"
Additional context to pass to the query analysis and qa steps
str
default:"''"
Filter the documents returned from the knowledge base
str
required
The text for semantic search. Note: you may add multiple upstream nodes to this field.
str
default:"''"
Filter the content returned from the knowledge base
bool
default:"False"
Whether to create a hybrid knowledge base
str
default:"'words'"
The method to break text into units before chunking. ‘words’: splits by word; ‘sentences’: splits by sentence boundary; ‘paragraphs’: splits by blank line/paragraph.
bool
default:"False"
Show intermediate steps
RetrievalConfig
RerankConfig
QueryEnhancementConfig
float
The alpha value for the retrieval
bool
Extract separate questions from the query and retrieve content separately for each question to improve search performance
bool
Do a natural language metadata query
bool
Expand query to improve semantic search
bool
Expand query terms to improve semantic search
int
The maximum number of relevant chunks to be returned
bool
Refine the initial ranking of returned chunks based on relevancy
str
Refine the initial ranking of returned chunks based on relevancy
str
The unit of retrieval. Chunks will return the most relevant chunks, Documents will return document metadata with snippets, and Pages will return complete pages with all chunks from pages containing relevant content
One of:
chunks, documents, pagesfloat
The score cutoff
bool
Transform the query for better semantic search
str
default:"'accurate'"
The mode to use for the advanced search
One of:
accurate, faststr
default:"'gpt-4o-mini'"
The model to use for the QA