LLM Operations
Batch and structured operations built on top of a language model.
Batch Run
Runs an LLM on each row of a DataFrame column. If no column is specified, all columns are used.
| Input | Type | Required |
|---|---|---|
| Table | other | Yes |
| watsonx API Endpoint | str | No |
| Column Name | str | No |
| Language Model | model | Yes |
| watsonx Project ID | str | No |
| Instructions | str | No |
Advanced options: API Key, Enable Metadata, Output Column Name
| Output | Type |
|---|---|
| LLM Results | Table |
Guardrails
Validates input text against multiple security and safety guardrails using LLM-based detection.
| Input | Type | Required |
|---|---|---|
| Guardrails | str | Yes |
| Input Text | str | Yes |
| Language Model | model | Yes |
Advanced options: API Key, Custom Guardrail Description, Enable Custom Guardrail, Heuristic Detection Threshold
| Output | Type |
|---|---|
| Pass | JSON |
| Fail | JSON |
LLM Selector
Routes the input to the most appropriate LLM based on OpenRouter model specifications
| Input | Type | Required |
|---|---|---|
| Judge LLM | other | Yes |
| Language Models | other | Yes |
| Input | str | Yes |
| Optimization | str | No |
Advanced options: Fallback to First Model, API Timeout, Use OpenRouter Specs
| Output | Type |
|---|---|
| Output | Message |
| Selected Model Info | Data, JSON |
| Routing Decision | Message |
Smart Router
Routes an input message using LLM-based categorization.
| Input | Type | Required |
|---|---|---|
| Input | str | Yes |
| Language Model | model | Yes |
| Routes | table | Yes |
Advanced options: API Key, Additional Instructions, Include Else Output, Override Output
Smart Transform
Uses an LLM to generate a function for filtering or transforming structured data and messages.
| Input | Type | Required |
|---|---|---|
| JSON | other | Yes |
| Instructions | str | Yes |
| Language Model | model | Yes |
Advanced options: API Key, Max Size, Sample Size
| Output | Type |
|---|---|
| Output | JSON |
| Output | Table |
| Output | Message |
Structured Output
Uses an LLM to generate structured data. Ideal for extraction and consistency.
| Input | Type | Required |
|---|---|---|
| Input Message | str | Yes |
| Language Model | model | Yes |
| Output Schema | table | Yes |
Advanced options: API Key, Schema Name, Format Instructions
| Output | Type |
|---|---|
| Structured Output | JSON |
| Structured Output | Table |
aGenerate
Generate mock data for user defined schema. If a dataframe is provided, the component will generate similar rows.
| Input | Type | Required |
|---|---|---|
| Input Table | other | No |
| Watsonx API Endpoint | str | No |
| Number of Rows to Generate | int | No |
| Language Model | model | Yes |
| Ollama API URL | str | No |
| Watsonx Project ID | str | No |
| Schema | table | No |
Advanced options: API Key, Instructions
| Output | Type |
|---|---|
| Output Table | Table |
aMap
Augment the input dataframe adding new columns defined in the input schema. Rows are processed independently and in parallel using LLMs.
| Input | Type | Required |
|---|---|---|
| Input Table | other | Yes |
| Watsonx API Endpoint | str | No |
| Instructions | str | No |
| Language Model | model | Yes |
| Ollama API URL | str | No |
| Watsonx Project ID | str | No |
| As List | bool | No |
| Schema | table | Yes |
Advanced options: API Key, Keep Source Columns
| Output | Type |
|---|---|
| Output Table | Table |
aReduce
Analyze the entire input dataframe at once and generate a new dataframe following the instruction and the required schema
| Input | Type | Required |
|---|---|---|
| Input Table | other | Yes |
| Watsonx API Endpoint | str | No |
| Instructions | str | No |
| Language Model | model | Yes |
| Ollama API URL | str | No |
| Watsonx Project ID | str | No |
| As List | bool | No |
| Schema | table | Yes |
Advanced options: API Key
| Output | Type |
|---|---|
| Output Table | Table |