dify - 💡(How to fix) Fix Add built-in output validation and auto-retry for LLM Node

Self Checks

• I have read the Contributing Guide and Language Policy.
• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report, otherwise it will be closed.
• Please do not modify this template :) and fill in all the required fields.

1. Is this request related to a challenge you're experiencing? Tell me about your story.

Currently, the LLM node does not come with native output validation capabilities. When I need to standardize the model's responses — such as checking content format, text length, required keywords or data types — I have to manually combine loop and conditional branch nodes to implement validation and retry logic. This approach makes the workflow structure bloated, increases the difficulty of building and maintaining pipelines. For practical scenarios like content generation, data warehousing and automated task execution, standardized and reliable output is essential. It would be much more convenient if the platform could provide native validation and retry features.

2. Additional context or comments

I hope to add configurable validation rules and automatic retry logic inside the LLM node, without relying on extra workflow nodes. Expected capabilities: Multiple validation rules for LLM outputs: Check whether the result contains specified text or keywords Restrict minimum and maximum text length Verify data types and standard format (e.g. complete JSON structure and required fields) Support regular expression matching Auto-retry after validation failure: Allow configuring the maximum number of retries Automatically feed back error prompts to the LLM, so the model can regenerate content that meets the requirements This feature will simplify workflow design, improve overall stability, and bring a better experience for business automation and content production scenarios.

3. Can you help us with this feature?

• I am interested in contributing to this feature.