Parameter Extractor

Node Overview

Node name: Parameter Extractor
Supported scenarios: Chatflow Agent (supported), Workflow Agent (supported)

Feature Description

The Parameter Extractor node uses a large language model to extract structured parameters from natural language text: configure the definitions of the parameters to extract (or import parameter definitions directly from a tool), and the model identifies and extracts the corresponding values from the input variable, outputting them for downstream nodes—most typically converting a user's natural language request into the call parameters needed by a downstream Tool node.

Node Display Content

1. Basic Information Display

• Icon: node-specific icon
• Name: node name (displayed as "Parameter Extractor" in the interface)
• Model in use: displayed the same way as the LLM node (empty when not configured)
• Note: explanatory text added by the user

2. Action Buttons

• More button: click to display the context menu content
• Run button: click to enter the test run interface

Node Edit Panel

1. Model Configuration

• Model selection: required, same as the LLM node's model configuration—parameter extraction is performed by this model

2. Input Variable

• Variable selection: type / to insert a variable and choose the source text to extract parameters from (such as Start/sys.query or an upstream node's output)

3. Extraction Parameters

• Add parameter: click + to manually define a parameter to extract (parameter name, type, description, required or not)
• Import from tool: import a tool's input parameter definitions as extraction targets in one click—typical usage: first use the Parameter Extractor to extract the parameters a tool needs from the user's natural language, then pass them to the downstream Tool node

4. Instruction

• Additional instruction: supplementary extraction instructions can be written (such as format requirements or ambiguity handling rules); the editor is the same as the LLM node's prompt input, supporting / to insert variables

5. Fault Tolerance Configuration

• Retry on failure: a switch; automatically retries when extraction fails
• Exception handling: a dropdown to select the exception handling strategy (defaults to None)

6. Next Step Configuration

• Adding a next node is supported

Execution Logic

The execution logic of the Parameter Extractor node is consistent with the Dify platform; the specific implementation is subject to the backend.

Context Menu Actions

The context menu contains the following options:

1. Change node: change the node type
2. Copy: copy the node
3. Duplicate: duplicate the node content
4. Delete: delete the node
5. Help link: jump to the help documentation
6. Run node: test-run the current node

Test Run

• Support status: test run supported
• Change node: changing the node type is supported
• Copy features: copy, cut, and delete operations are supported

Use Cases

1. Natural Language to Tool Parameters (Most Typical)

• The user says "Show me Tesla news from the last two weeks" → extract symbol="TSLA.US", days=14 → pass them to the downstream Tool node
• Combined with Import from tool: use the tool's input parameter definitions directly as extraction targets, with no need to build parameters by hand

2. Structured Information Collection

• Extract structured fields such as names, dates, and amounts from the user's natural language or multi-turn conversation text
• The extraction results can drive IF Else branching, be processed by a Code node, or be referenced by an Answer node

3. Ambiguity and Default Value Handling

• Define ambiguity handling rules in the Instruction (e.g., "default to US stocks when the market is unspecified", "convert relative date expressions to absolute dates")

FAQ

Q: What if extraction is inaccurate?

A: Make the parameter's description more specific (including format and examples), add rules in the instruction, or switch to a more capable model (see Tip 2 in section 2.5).

Q: How do I combine it with the Tool node?

A: Use Import from tool to import the tool's input parameter definitions in one click; the fields of the extraction result can then plug directly into the downstream Tool node.

Q: What if extraction fails or fields are missing?

A: Enable Retry on failure and configure the Exception handling strategy according to the importance of the flow; for required parameters, add fallback checks downstream.

Considerations

1. Model is required: extraction is performed by the selected model, and extraction quality depends on the model's capability
2. Parameter description quality: the clearer the parameter names, types, and descriptions, the more accurate the extraction
3. Fault tolerance configuration: for important flows, be sure to configure retry on failure and exception handling
4. Source text: the input variable should point to text containing the target information (such as sys.query or an upstream output)