HuanLinOTO/dsh-plugin-mineru
251Last commit Aug 15, 2026
dsh-plugin-mineru DSH plugin
This plugin integrates MinerU, a document parsing service, into DSH. It provides five tools for synchronous parsing, asynchronous job submission, status polling, result retrieval, and health checks. Users must configure the MinerU API endpoint and optionally set an API key.
How to install the dsh-plugin-mineru DSH plugin
dsh plugin --profile web add @huanlin/dsh-plugin-mineruCopying does not run this command. Review the repository and version before installing the dsh-plugin-mineru DSH plugin.
dsh-plugin-mineru DSH plugin data source
dsh-plugin-mineru DSH plugin snapshot date: Aug 16, 2026
discovered
What the dsh-plugin-mineru DSH plugin can do
- Synchronously parse local documents and return extracted Markdown
- Asynchronously submit parse jobs and retrieve task IDs
- Poll task status (pending/processing/completed/failed)
- Retrieve completed task results, with fallback to file storage for large outputs
- Check MinerU server health, version, queue depth, and concurrency
- Support multiple file formats: PDF, images, DOCX, PPTX, XLSX
- Configurable backend (pipeline, vlm-engine, hybrid-engine, etc.) and parse method (auto, txt, ocr)
Where the dsh-plugin-mineru DSH plugin fits
- Enable AI models to read and analyze PDF documents or scanned images
- Automate batch conversion of Office files to Markdown for downstream processing
- Extract text from images and tables using OCR or VLM engines
- Integrate document parsing into agent workflows that require structured content
- Monitor MinerU server status and capacity from within DSH
Who the dsh-plugin-mineru DSH plugin is for
- Developers who deploy MinerU and want to expose its parsing capabilities to DSH-powered models
- Users who need to convert diverse document formats into machine-readable Markdown/JSON
dsh-plugin-mineru DSH plugin limitations
- Requires a running MinerU instance; the plugin does not include or deploy MinerU itself
- Inline Markdown output is capped at 200,000 characters; larger results are saved to temporary files
- Depends on network connectivity to the MinerU API server; proper configuration of baseURL and optional API key is mandatory
- Asynchronous polling has a timeout of 10 minutes (configurable) for long documents
dsh-plugin-mineru DSH plugin: from the repository README
Quoted from the HuanLinOTO/dsh-plugin-mineru README, the upstream source of the dsh-plugin-mineru DSH plugin. Copyright remains with the original authors.
> 📌 本插件已收录于 [dshfind](https://dshfind.com/zh) 插件超市，点击上方徽章直达主页。 # dsh-mineru [](https://www.npmjs.com/package/@huanlin/dsh-plugin-mineru) DSH 插件：向模型暴露 [MinerU](https://github.com/opendatalab/MinerU) 文档解析工具。MinerU 可将 PDF、图片、DOCX、PPTX、XLSX 等文件转换为结构化的 Markdown / JSON。 ## 安装 ```sh # 从 npm 安装（推荐）： dsh plugin --profile web add @huanlin/dsh-plugin-mineru # 从本地 checkout 开发安装： dsh plugin --profile web add link:D:\Projects\deepseek-harness\dsh-mineru ``` 若从 git 安装（pnpm ≥10），需在 profile 的 `pnpm-workspace.yaml` 中允许构建： ```yaml allowBuilds: '@huanlin/dsh-plugin-mineru': true ``` ## 配置 在 DSH GUI 设置页或 `cordis.patch.yml` 中配置： | 字段 | 类型 | 默认值 | 说明 | |------|------|--------|------| | `baseURL` | string | *(必填)* | MinerU API 地址，如 `http://your-mineru-host:18000` | | `apiKeyEnv` | credential-ref | `MINERU_API_KEY` | API key 的环境变量名 / 凭据引用。测试实例无需鉴权。 | | `defaultBackend` | enum | `pipeline` | `pipeline` / `vlm-engine` / `hybrid-engine` / `vlm-http-client` / `hybrid-http-client` | | `defaultParseMethod` | enum | `auto` | `auto` / `txt` / `ocr` | | `defaultLang` | string | `ch` | pipeline 后端的语言代码 | | `pollIntervalMs` | number
Read the full READMERepository license: NOASSERTION
dsh-plugin-mineru DSH plugin questions
How do I install dsh-plugin-mineru?
Run `dsh plugin --profile web add @huanlin/dsh-plugin-mineru` in your terminal. If you are installing from a local checkout, use `dsh plugin --profile web add link:/path/to/dsh-mineru`. For git installations with pnpm ≥10, you need to allow build in the profile's `pnpm-workspace.yaml`.
How do I configure the MinerU API address?
In the DSH GUI settings or in `cordis.patch.yml`, set the `baseURL` field to your MinerU API endpoint, e.g., `http://your-mineru-host:18000`. If your MinerU instance requires authentication, also set the `apiKeyEnv` field to reference an environment variable containing the API key.
Why is the parsed Markdown truncated?
The plugin has a configurable maximum character limit for inline Markdown output, defaulting to 200,000 characters. If the parsed content exceeds this limit, the full result is saved to a temporary file, and the model receives the truncated version. You can increase the `maxMdOutputChars` setting, but be aware of memory and performance implications.
Can I parse images or scanned PDFs with this plugin?
Yes, MinerU supports image files and scanned PDFs. The plugin can use different backends (pipeline, vlm-engine, hybrid-engine) and parse methods (auto, txt, ocr). For OCR-based extraction, set the `defaultParseMethod` to `ocr` or use the appropriate backend. The `vlm-engine` backend is particularly suited for visual document understanding.
How do I handle large documents or batch processing?
For large documents, use the asynchronous tools: `mineru_submit_parse_job` to submit the task, `mineru_get_parse_status` to poll until completion, and `mineru_get_parse_result` to retrieve the result. This avoids blocking the model and allows parallel submission of multiple files. The polling timeout is 10 minutes by default, adjustable via `pollTimeoutMs`.