Skip to content
FlowAI Logo FlowAI 文档

Text Truncate Node

The Text Truncate node shortens long text to a fixed length. It supports character-based and line-based truncation, making it ideal for keeping LLM replies concise, summarizing logs, or preparing snippets for downstream steps.

Node Configuration

Basic Settings

text truncate node

  1. Node Name

    • Use an action-oriented name like “Trim AI Reply”
    • Output variables inherit the node name prefix, for example $Trim AI Reply.text

  2. Text Source

    • Choose a variable from upstream nodes such as $LLM.output or $input.content
    • Any context variable reference is supported

  3. Mode

    • Characters (default): Truncates based on Unicode character count
    • Lines: Splits by newline \n and keeps the first N lines

  4. Limit

    • Accepts zero or any non-negative integer
    • A value of 0 outputs an empty string
    • The field is required and supports variable references (for example $settings.limit)

  5. Output Variable

    • Defaults to text
    • Downstream nodes read the result using $NodeName.text

How It Works

  • During execution, the node resolves $node.field placeholders across all inputs, including nested paths.
  • If the limit is empty or invalid, the workflow throws an error to avoid ambiguous results.
  • The implements of the truncation logic:
    • Characters: Truncates based on Unicode character count, supports emoji, Chinese, etc.
    • Lines: Splits on newlines and joins back the first N lines
  • After truncation, the node writes the result into context, visible in debug outputs under $NodeName.text.

Usage Examples

1. Cap LLM Responses (Characters)

  1. Insert a Text Truncate node after your LLM node.
  2. Configure it as follows: 
    Text: $LLM.output
    Mode: Characters
    Limit: 280
  3. Pass $Text Truncate.text to a Content Combiner or Output node to ensure the reply stays within 280 characters.

2. Show the First Lines of Logs (Lines)

  1. Fetch multi-line logs via an Input or HTTP node.
  2. Configure the Text Truncate node: 
    Text: $Fetch Logs.response
    Mode: Lines
    Limit: 5
  3. Reference $Text Truncate.text downstream to display only the first five lines.

Advanced Techniques

  • Dynamic Limits: Drive the limit with environment-specific variables, such as $Config.maxChars, for flexible policies.
  • Chained Control: Combine line truncation followed by character truncation when you need structural and length guarantees.
  • Fallback Handling: Pair with conditional nodes to provide default messages when truncation yields an empty result.

Best Practices

  • Preview Input Size: Use the debugger or Content Combiner to confirm the original length before truncating.
  • Clarify Units: Document whether teammates should configure characters or lines to avoid misconfiguration.
  • Preprocess Large Text: For extremely long content, preprocess with other nodes (JSON extraction, code execution) before truncating.
  • Idempotent Output: Ensure truncated results remain meaningful if the workflow runs multiple times on the same data.

Common Issues