feat: add Claude Opus 5 to LDX hub plugin and bump to v0.0.20

[HidekiMori]

LDX hub for Dify is a document AI gateway plugin that brings two capabilities into Dify:

• StructFlow: Extract structured JSON from unstructured text using leading LLMs
• RefineLoop: Iterative XLIFF translation refinement

Supports OpenAI, Microsoft Azure OpenAI, Google, Anthropic, Amazon Bedrock, and xAI through a single API and a single billing system.

Plugin source: https://github.com/ldxhub-io/dify-nodes-ldxhub

Plugin Submission Form

1. Metadata

• Plugin Author: ldxhub-io
• Plugin Name: ldxhub
• Repository URL: https://github.com/ldxhub-io/dify-nodes-ldxhub

2. Submission Type

• New plugin submission
• Version update for existing plugin

3. Description

LDX hub for Dify is a document AI gateway plugin that unifies access to leading LLM providers (OpenAI, Microsoft Azure OpenAI, Google, Anthropic, Amazon Bedrock, xAI) through a single API and billing system.

The plugin provides two tools:

StructFlow — Extract structured JSON from unstructured text. Define an extraction schema with a system prompt and an example output, then process JSONL records through the model of your choice. Output is a JSONL of structured data per record. Useful for processing medical records, customer feedback, legal documents, financial reports, and other unstructured text at scale.

RefineLoop — Iteratively improve XLIFF translation quality. Built on the same engine as StructFlow, specialized for translation review. Each segment goes through multiple revision rounds where the AI critiques and improves the translation, with structured revision notes. XLIFF tag integrity is validated after every revision. Real-world benchmark: BERT paper translation (65,000 source characters, 145,000 with tags) refined with Gemini 3.5 Flash in 4 minutes 9 seconds end-to-end on Dify Cloud.

Both tools support an optional webhook_url parameter for long-running jobs.

The plugin supports bilingual README (English / Japanese) via Dify's multilingual README feature.

What's new in v0.0.20

This version adds support for Anthropic's Claude Opus 5 to both StructFlow and RefineLoop tools. Claude Opus 5 is the Opus-tier model of the Claude 5 generation, offering Claude Fable 5-class quality at half the token cost.

No breaking changes. Existing workflows continue to work as before.

Risk Level

• Low risk
• Medium risk
• High risk

This is a metadata-only change: it adds one model option to the existing dropdowns. No code, dependency, or behavioral changes. The plugin performs none of the sensitive operations listed in the submission requirements (no command/code execution, no database or filesystem access, no arbitrary URL fetching, no sensitive personal data processing).

4. Checklist

• I have read and followed the Publish to Dify Marketplace guidelines
• I have read and comply with the Plugin Developer Agreement
• I confirm my plugin works properly on both Dify Community Edition and Cloud Version
• I confirm my plugin has been thoroughly tested for completeness and functionality
• My plugin brings new value to Dify

5. Documentation Checklist

Please confirm that your plugin README includes all necessary information:

• Step-by-step setup instructions
• Detailed usage instructions
• All required APIs and credentials are clearly listed
• Connection requirements and configuration details
• Link to the repository for the plugin source code

6. Privacy Protection Information

Based on Dify Plugin Privacy Protection Guidelines:

Data Collection

This plugin does not collect any user personal data within the plugin itself. It only forwards user-provided data (text, JSONL files, XLIFF files) to LDX hub servers and the configured AI provider for processing. The plugin itself stores no data and sends no telemetry.

User-provided data sent to third parties:

• LDX hub servers (https://gw.ldxhub.io) — for document processing
• Configured AI provider (OpenAI, Microsoft Azure OpenAI, Google, Anthropic, Amazon Bedrock, or xAI) — selected by the user via the model parameter

API keys are encrypted by Dify (PKCS1_OAEP) and stored within the user's workspace.

Privacy Policy

• I confirm that I have prepared and included a privacy policy in my plugin package based on the Plugin Privacy Protection Guidelines

Privacy policy: https://github.com/ldxhub-io/dify-nodes-ldxhub/blob/main/PRIVACY.md

[crazywoola]

✅ LGTM

Decision: Approve

Local Check Results

Check	Status	Detail	Required action
PR content language	✅ Pass	PR title/body CJK ratio=0.0% (zh=0, en=3587, ignored_zh=0, allowed_zh<=0)	None.
Project structure	✅ Pass	All expected files present: manifest.yaml, README.md, PRIVACY.md. _assets/: yes.	None.
Manifest author	✅ Pass	author is valid.	None.
Icon validation	✅ Pass	icon exists: _assets/icon.svg	None.
Version check	✅ Pass	version 0.0.20 is available.	None.
README language	✅ Pass	README.md CJK ratio=0.0% (zh=0, en=5749, allowed_zh<=0)	None.
PRIVACY.md	✅ Pass	PRIVACY.md exists and is non-empty.	None.
Dependency install	✅ Pass	requirements installed successfully.	None.
dify_plugin version	✅ Pass	dify_plugin version 0.9.1 satisfies >= 0.9.0.	None.
Install test	✅ Pass	plugin install test passed.	None.
Packaging test	✅ Pass	packaging check passed.	None.