Hub
Pricing About
Collection

KNIME for Generative AI

AILLMChatModelEmbeddings
+2

Explore this collection of ready to use workflows to get started with using Large Language Models (LLMs). Browse workflows for connecting to and interacting with LLMs, building chat apps, performing retrieval-augmented generation (RAG), detecting vulnerabilities in LLMs and RAG-based systems and more.

Want to learn how to integrate GenAI in your data workflows?
Take our free self-paced course that will introduce you to the nodes needed to start enriching analytics with LLMs and GenAI in KNIME.
Jul 30, 2025 2:35 PM
Like

Connect to AI providers

In this section, we introduce the nodes needed to authenticate with different AI providers and select the preferred LLM. The KNIME AI extension provides built-in connectivity to different AI providers, such as OpenAI (and Azure OpenAI), Anthropic, Gemini via Google AI Studio and Vertex AI, DeepSeek, Databricks AI, IBM watsonx.ai, Hugging Face and GPT4All.
OpenAI
OpenAI is most known for its application ChatGPT and offers some of the most state-of-the-art general-purpose LLMs, such as GPT-4o and beyond. These models are designed to handle a wide range of natural language processing tasks with strong performance and versatility. OpenAI also supports vision models with image generation and editing capabilities, allowing users to create and modify visuals from natural language prompts.
Node / Source
OpenAI Authenticator

Validates OpenAI API keys by sending a request to the OpenAI models endpoint without consuming tokens. Learn more

Node / Source
OpenAI LLM Selector

Establishes a connection with an OpenAI LLM, allowing selection from a predefined or comprehensive list of available models. Learn more

Node / Visualizer
OpenAI Image Generator

Generates or edits images based on prompts using OpenAI’s DALL-E 3 or GPT Image 1 models, supporting PNG format. Learn more

Go to item
Workflow
OpenAI
Demonstrates prompt-based LLM calls in KNIME using the OpenAI integration: send prompts, receive responses, and manage authentica…
AI Extension Guide1. Prompting LLMProvidersOpenAI
0
knime
Go to item
Workflow
Image generation and editing with GenAI
Image generationImage editingOpenAI
+2
Uses KNIME’s OpenAI Image Generator node to generate or edit images via DALL·E‑3 or GPT Image 1—supports loops, batch output, and…
AI Extension Example Workflows5) Use CasesImage generation and editing with GenAI
0
knime
Anthropic
Anthropic develops the Claude series of LLMs, focused on safety, reliability, and helpfulness in a wide range of conversational and reasoning tasks.
Node / Source
Anthropic Authenticator

Authenticates with the Anthropic API using an API key for secure access. Learn more

Node / Source
Anthropic LLM Selector

The item connects to an Anthropic LLM, allowing users to select a model after authentication. Learn more

Go to item
Workflow
Anthropic
Demonstrates integration of Anthropic models into KNIME workflows using dedicated Anthropic‑node authentication and prompt execut…
AI Extension Guide1. Prompting LLMProvidersAnthropic
0
knime
Google Gemini
Gemini is a family of multimodal LLMs developed by Google DeepMind, designed to handle text, code, and image inputs for a wide range of tasks, available from Google AI Studio or via Google Cloud’s Vertex AI platform.
Node / Source
Google AI Studio Authenticator

Authenticates with Google AI Studio using an API key for selecting language and embedding models from the Gemini family. Learn more

Node / Source
Gemini LLM Selector

Facilitates the selection of a Gemini language model using an authenticated connection from Google services. Learn more

Node / Source
Vertex AI Connector

Connects to a Vertex AI project in Google Cloud for selecting chat and embedding models from the Gemini family. Learn more

Go to item
Workflow
Gemini
Demonstrates use of Google AI Studio and Gemini LLM/embedding nodes in KNIME workflows, including model selection and prompt inte…
AI Extension Guide1. Prompting LLMProvidersGemini
0
knime
Go to item
Workflow
Vertex AI
Connects KNIME to Google Cloud Vertex AI, enabling use of Gemini chat and embedding models via the Vertex AI Connector node.
AI Extension Guide1. Prompting LLMProvidersVertex AI
0
knime
Databricks AI
Databricks AI offers a suite of artificial intelligence capabilities and features integrated within the Databricks Data Intelligence Platform for managing and deploying generative AI models, including proprietary models like DBRX and support for open-source and third-party foundation models.
Node / Source
Databricks Workspace Connector

Connects to Databricks or Azure Databricks workspaces, enabling data integration and processing within those environments. Learn more

Node / Source
Databricks LLM Selector

Connects to a Large Language Model served by a Databricks workspace for processing language tasks. Learn more

Go to item
Workflow
Databricks AI
DatabricksLLMVector Store
+1
This workflow shows how to use Databricks nodes in KNIME. It is part of the KNIME AI Extension Guide.
AI Extension Guide1. Prompting LLMProvidersDatabricks AI
0
knime
GPT4All
GPT4All is an open-source AI platform by Nomic AI that enables running language models locally without internet access. It supports lightweight models optimized for privacy and offline use.
Node / Source
Local GPT4All LLM Selector

Connects to a local GPT4All language model for processing natural language tasks. Learn more

Go to item
Workflow
GPT4All
LLMAIGPT4All
Local LLM integration using GPT4All: install GPT4All, download a GGUF model, and connect via the GPT4All LLM Connector to prompt …
AI Extension Guide4. AI GovernanceGPT4All
0
knime

Prompt LLMs

In this section, we introduce the nodes that can be used to craft text instructions, and the prompting nodes that send the instructions to the selected LLM.
Prompt engineering
In KNIME, it’s possible to craft prompts as a simple string of text or combining text with images.
Node / Manipulator
Expression

The item enables row-by-row data manipulation and transformation within workflows, allowing for complex expressions and calculations. Learn more

Node / Manipulator
Message Creator

Generates messages with text and/or image content, assigning roles like User, AI, or Tool, and supports embedding tool calls. Learn more

Node / Source
Table Creator

Allows the manual creation of a data table. The data can entered in a spreadsheet like table. Learn more

Node / Manipulator
String Manipulation

Manipulates strings by performing operations like search and replace, capitalization, or trimming whitespace. Learn more

LLM Prompters
The KNIME AI Extension supports sending prompts to an LLM in two fashions: one-shot prompts for each input row without retaining any conversation history, or chat-style prompts where the conversation history is input to generate context-aware responses.
Node / Predictor
LLM Prompter

Processes each row in an input table by sending a prompt to an LLM and receiving an isolated response for each. Learn more

Node / Predictor
LLM Chat Prompter

The item generates AI responses by using conversation history and optional tool definitions to enhance chat interactions. Learn more

Go to item
Workflow
LLM Prompter
Part of the AI Extension Guide; demonstrates use of the LLM Prompter node for prompt‑based LLM interaction in KNIME, including au…
AI Extension Guide1. Prompting LLMLLM Prompter
0
knime
Go to item
Workflow
LLM Chat Prompter
Demonstrates chat-based prompting using the Chat LLM Prompter node—it supports multi-turn conversation history and evaluation wor…
AI Extension Guide1. Prompting LLMLLM Chat Prompter
0
knime

Build vector stores and RAG systems

KNIME makes it easy to enhance AI applications with custom data. The KNIME AI Extension supports creating and managing vector stores, which are commonly used in retrieval-augmented generation (RAG) systems. RAG systems leverage user-curated, custom knowledge bases to help LLMs generate domain-specific responses while reducing the risk of hallucinations.
Vector store creation and management
Vector stores are specialized databases that store and manage objects—such as documents, code, or dictionaries—as vectors in a multidimensional space, also known as embeddings. The KNIME AI Extension includes nodes to preprocess documents, generate embeddings using various AI providers, and create and manage vector stores.
Node / Manipulator
Text Chunker

Splits large texts into smaller overlapping chunks while maintaining semantic or syntax coherence. Learn more

Node / Source
FAISS Vector Store Creator

Generates a FAISS vector store to map documents into numerical vectors capturing their semantic meaning using an embedding model. Learn more

Node / Other
Vector Store Data Extractor

Extracts documents, embeddings, and metadata from a vector store into a table for merging with other vector stores. Learn more

Node / Source
OpenAI Authenticator

Validates OpenAI API keys by sending a request to the OpenAI models endpoint without consuming tokens. Learn more

Node / Source
OpenAI Embedding Model Selector

The item connects to an OpenAI Embedding Model, allowing users to select and use available embedding models after authentication. Learn more

Node / Predictor
Text Embedder

Generates dense vector representations of text to capture semantic meaning for similarity comparisons in a high-dimensional space. Learn more

Go to item
Workflow
Create a vector store
OpenAIEmbeddingsVector Store
+1
Uses OpenAI embeddings (FAISS) to build a vector store from KNIME table data and enables similarity search via Vector Store Retri…
AI Extension Example Workflows3) Vector Stores3.1 - Create a Vector Store
2
knime
Go to item
Workflow
Update a vector store
LLMVector StoreAI
+3
Demonstrates updating an existing vector store using local embedding updates or adding new documents (incremental storage).
EducationGetting Started CoursesAI Chatbots, RAG & Governance with Data Workflows - Getting Started03 Update a vector store
0
knime
Context-aware responses with RAG
After building vector stores with custom knowledge bases, they can be queried to retrieve semantically relevant documents (Retrieval). This retrieved information enriches the user prompt with contextual details (Augmented), enabling the LLM to generate more accurate, relevant or domain-specific responses (Generation).
Node / Source
Vector Store Retriever

Retrieves relevant embeddings from a vector store based on user queries for similarity search. Learn more

Node / Manipulator
Expression

The item enables row-by-row data manipulation and transformation within workflows, allowing for complex expressions and calculations. Learn more

Node / Source
OpenAI LLM Selector

Establishes a connection with an OpenAI LLM, allowing selection from a predefined or comprehensive list of available models. Learn more

Node / Predictor
LLM Prompter

Processes each row in an input table by sending a prompt to an LLM and receiving an isolated response for each. Learn more

Go to item
Workflow
Retrieval‑Augmented Generation (RAG)
LLMVector StoreAI
+3
Demonstrates a visual RAG implementation in KNIME: creating vector stores, retrieving relevant context, and augmenting prompts fo…
EducationGetting Started CoursesAI Chatbots, RAG & Governance with Data Workflows - Getting Started02 Retrieval-Augmented Generation (RAG)
0
knime
Go to item
Workflow
Product FAQ using RAG
OpenAIFAISSVector Store
+1
Uses Retrieval‑Augmented Generation on a FAQ dataset: embedding product Q&A, creating vector store, retrieving relevant entries, …
AI Extension Guide2. RAGProduct FAQ
0
knime

Build AI chatbots

KNIME simplifies the creation of AI-driven chatbots that incorporate conversation history and an interactive chat view.
Node / Manipulator
Message Creator

Generates messages with text and/or image content, assigning roles like User, AI, or Tool, and supports embedding tool calls. Learn more

Node / Source
OpenAI Authenticator

Validates OpenAI API keys by sending a request to the OpenAI models endpoint without consuming tokens. Learn more

Node / Source
OpenAI LLM Selector

Establishes a connection with an OpenAI LLM, allowing selection from a predefined or comprehensive list of available models. Learn more

Node / Predictor
LLM Chat Prompter

The item generates AI responses by using conversation history and optional tool definitions to enhance chat interactions. Learn more

Node / Visualizer
Agent Chat View

Facilitates interactive, multi-turn conversations with an AI agent, using tools and input data to provide context-aware responses. Learn more

Go to item
Workflow
AI Chatbots
AIGenAIAI chatbot
+4
Demonstrates how to build custom chatbots using KNIME: from single Q&A bots to full interactive chat components with conversation…
EducationGetting Started CoursesAI Chatbots, RAG & Governance with Data Workflows - Getting Started01 AI Chatbots
1
knime

AI Governance

The KNIME AI Extension and KNIME Software offer dedicated nodes and features to ensure the safe and responsible development and deployment of GenAI technologies. This includes a suite of governance features and safeguarding mechanisms to evaluate LLMs and RAG-based systems, centrally manage and control access to AI models, and automates the end-to-end deployment of AI solutions safely into production.
Evaluate vulnerabilities in LLMs and RAG-based systems
AI vulnerabilities are weaknesses and risks in LLMs or RAG-based systems that can be exploited to produce incorrect, biased, or harmful outputs, or pose a severe security threat for sensitive information. The KNIME AI Extension supports dedicated nodes that integrate Giskard’s capabilities for testing, monitoring, and evaluating AI models.
Node / Other
Giskard LLM Scanner

Evaluates workflows for vulnerabilities in GenAI models using heuristics and LLM-assisted detectors to identify issues like misinformation and harmful content. Learn more

Node / Other
Giskard RAGET Test Set Generator

Generates diverse test questions to evaluate and improve the performance of various components of the RAG system. Learn more

Node / Other
Giskard RAGET Evaluator

Evaluates the correctness of a RAG system’s answers by comparing them to reference answers using an LLM. Learn more

Node / ScopeStart
Capture Workflow Start

Marks the beginning of a workflow segment to be captured and made available at the output of a corresponding end node. Learn more

Node / ScopeEnd
Capture Workflow End

Marks the end of a captured workflow segment, making the entire segment available at the workflow output port. Learn more

Go to item
Workflow
Evaluate an LLM-based application
GenAIGiskardLLM
+2
Runs Giskard’s LLM Scanner node to automatically detect vulnerabilities—like hallucination, bias, prompt injection, harmful conte…
AI Extension Example Workflows6) EvaluationEvaluation of Text Generation Workflows with Giskard
0
knime
Go to item
Workflow
Evaluate a RAG-based system
GenAIGiskardRAG
+2
Uses Giskard RAGET toolkit to generate diverse test questions and evaluate a RAG system’s retrieval and generation quality in KNI…
AI Extension Example Workflows6) EvaluationEvaluation of RAG Workflows with Giskard RAGET
1
knime
Centralized model access and management
For enterprise users interested in centralized model governance and access control, KNIME Business Hub includes GenAI Gateway. This feature allows admins to configure and manage LLMs and embedding models centrally across the organization.
Node / Source
KNIME Hub Authenticator

Authenticates against a KNIME Hub to enable compatible nodes to access the configured Hub. Learn more

Node / Source
KNIME Hub LLM Selector

Connects to a Large Language Model to generate text, answer questions, summarize content, or perform other text-based tasks. Learn more

Node / Source
KNIME Hub Embedding Model Selector

Generates dense vector representations of text input data for tasks like similarity search, clustering, and classification. Learn more

Node / Source
KNIME Hub AI Model Lister

Lists available models in the GenAI Gateway of the connected KNIME Hub, providing their name, type, and description. Learn more

Go to item
Workflow
Meeting transcript generation and summarization with KNIME GenAI Gateway
GenAIPrompt EngineeringGenAI Gateway
+4
Connects to GenAI Gateway to generate fictional meeting transcripts, then uses a chat‑model (e.g. GPT‑4o) to extract summaries, a…
AI Extension Example Workflows5) Use CasesMeeting transcript generation and summarization with KNIME GenAI Gateway
0
knime
Go to item
Workflow
CV summarization with KNIME GenAI Gateway and local LLMs
AILLMSummarization
+4
Summarizes applicants’ CVs using GenAI Gateway and local LLMs, with PII anonymization and interactive dashboard for HR screening.
AI Extension Example Workflows5) Use CasesGenAI Summarizer of job applicants' CVs
1
knime
Continuous Deployment for Data Science (CDDS) for Business Hub
The Continuous Deployment for Data Science (CDDS) extension for KNIME Business Hub automates the end-to-end process of deploying data science and AI solutions safely into production. It offers users an intuitive UI for deployment, validation, monitoring, and retraining, while administrators can oversee the entire deployment flow.
Go to item
Collection
Getting started with Continuous Deployment for Data Science (CDDS) for Business Hub
CDDSMLOpsDevOps
+5
Browse example workflows, node explanations and further resources for learning about CDDS for Business Hub.
1
knime

Use cases

AI can be applied in various industries and for different tasks, such as marketing campaigns and financial services. Here are some common use cases that utilize KNIME AI capabilities.
Go to item
Workflow
City and neighborhood insights with Geo-GenAI
GenAIGeospatial analyticsPublic sector
+2
Explores Zurich city data with maps, charts, and LLM-generated insights using GenAI.
AI Extension Example Workflows5) Use CasesCity and neighborhood insights with Geo-GenAI
2
knime
Go to item
Workflow
Feature Engineering with GenAI for Classification
Machine learningLLMsGenAI
+4
Generates ML-ready features from text using GenAI for loan approval classification.
AI Extension Example Workflows5) Use CasesFeature Engineering with GenAI for Classification
5
knime
Go to item
Workflow
Research Paper Summaries Generation using Generative AI
Release 5.3AIText Chunk
+1
Summarizes long medical papers using LLM chunking and prompt-based generation.
AI Extension Example Workflows5) Use CasesResearch Paper Summaries Generation using Generative AI
4
knime
Go to item
Workflow
Google Ads Assets Creation with Generative AI
Release 5.3Google AdsKeyword Research
+2
Creates ad headlines by combining keyword insights with GenAI prompting.
AI Extension Example Workflows5) Use CasesGoogle Ads Assets Creation with Generative AI
1
knime
Go to item
Workflow
Explain KPI report with multimodal LLMs
LLMsGenAIKPI report
+3
Analyzes KPI charts with local vision-language models to generate narrative explanations.
AI Extension Example Workflows5) Use CasesExplain KPI report with multimodal LLMs
2
knime

Learning Resources

Learn how to integrate Generative AI and data workflows, build AI chatbots and custom RAG systems, and apply AI governance best practices using KNIME. Take these free, self-paced courses and earn a microcredential on completion.
Online free course: GenAI & Data Workflows – Getting Started
Online free course: AI Chatbots, RAG, & Governance with Data Workflows – Getting Started

Additional Material

This section includes additional resources and materials to complement the main content.
Docs: KNIME AI Extension Guide
Blog: An introduction to Large Language Models
Blog: How to Classify Sentiment with Generative AI
Blog: What are AI hallucinations & how to mitigate them in LLMs
Blog: Mitigate hallucinations with retrieval augmented generation in KNIME
Blog: Evaluate LLM, RAG, and ML vulnerabilities in KNIME with new Giskard nodes
Use cases: Summarize with GenAI

KNIME
Open for Innovation

KNIME AG
Talacker 50
8001 Zurich, Switzerland
  • Software
  • Getting started
  • Documentation
  • Courses + Certification
  • Solutions
  • KNIME Hub
  • KNIME Forum
  • Blog
  • Events
  • Partner
  • Developers
  • KNIME Home
  • Careers
  • Contact us
Download KNIME Analytics Platform Read more about KNIME Business Hub
© 2025 KNIME AG. All rights reserved.
  • Trademarks
  • Imprint
  • Privacy
  • Terms & Conditions
  • Data Processing Agreement
  • Credits