A native dsh plugin that wires MinerU document parsing into DeepSeek Harness: submit a local file and get structured Markdown / JSON back through model-callable tools.
DSH integration
Native runtime
Author-claimed
Safety audit
Unaudited
Last verified
2026-09-01
License
NOASSERTION
01What can it help you accomplish?
Let the model parse PDF, images, DOCX, PPTX and XLSX into structured Markdown / JSON inside DeepSeek Harness
Structured Markdown or JSON extracted from the submitted document, returned inline to the model
DeepSeek Harness (dsh) users who need document parsing and knowledge extraction without leaving the agent
02How to install into DeepSeek Harness
Prerequisites
- A running MinerU API host reachable from DSH (baseURL is required, e.g. http://your-mineru-host:18000)
- Official dsh CLI available; for git installs with pnpm >=10 the build must be allowed in pnpm-workspace.yaml
Installation steps
- 01
Install from npm (recommended): `dsh plugin --profile web add @huanlin/dsh-plugin-mineru`
$ dsh plugin --profile web add @huanlin/dsh-plugin-mineru
- 02
Or link a local checkout for development: `dsh plugin --profile web add link:<path-to-dsh-mineru>`
$ dsh plugin --profile web add link:<path-to-dsh-mineru>
Verify the integration
- Run `dsh plugin --profile web list` and confirm @huanlin/dsh-plugin-mineru appears among installed plugins
Rollback
- Uninstall with `dsh plugin --profile web remove @huanlin/dsh-plugin-mineru`
03DSH integration and capability boundaries
Native dsh plugin installed via `dsh plugin --profile web add @huanlin/dsh-plugin-mineru`; exposes MinerU document-parsing tools to the model
mineru_parse_document
local document file (PDF / image / DOCX / PPTX / XLSX)→extracted markdown text returned to the model
when output exceeds maxMdOutputChars (default 200000) the full content is written to a temp filemineru_submit_parse_job
local document file for asynchronous parsing→task_id returned immediately
mineru_get_parse_status
task_id of a submitted parse job→status: pending / processing / completed / failed
mineru_get_parse_result
task_id of a completed parse job→extracted markdown / JSON result of the parse job
04Who is it for? When not to use it?
Good for
- DeepSeek Harness (dsh) users who need document parsing and knowledge extraction without leaving the agent
Not for
- baseURL is required and must point to a reachable MinerU API host (e.g. http://your-mineru-host:18000). A test instance may run without authentication, but the endpoint must be configured before use.
05Compatibility, maintenance and safety notes
- baseURL is required and must point to a reachable MinerU API host (e.g. http://your-mineru-host:18000). A test instance may run without authentication, but the endpoint must be configured before use.
- When installing from git with pnpm >=10 you must allow the build in your profile's pnpm-workspace.yaml (`allowBuilds: '@huanlin/dsh-plugin-mineru': true`), otherwise the install will not build.
- Inline markdown returned to the model is capped at maxMdOutputChars (default 200000); content beyond the limit is stored in a temp file instead of being returned inline.
NOASSERTION license · single maintainer (huanlinoto) · no tagged release recorded (latest_release is null)
06Frequently asked questions
How do I install dsh-plugin-mineru?
Install it with the official dsh CLI: `dsh plugin --profile web add @huanlin/dsh-plugin-mineru`. For local development you can link a checkout with `dsh plugin --profile web add link:<path>`. If you install from git with pnpm >=10, allow the build in your profile's `pnpm-workspace.yaml`.
What does it do?
It exposes MinerU document-parsing tools to the model so it can turn PDF, images, DOCX, PPTX and XLSX files into structured Markdown / JSON. The recommended `mineru_parse_document` tool submits a file, polls until done, and returns the markdown automatically.
What configuration is required?
You must set `baseURL` to your MinerU API host (e.g. `http://your-mineru-host:18000`). Other options — `apiKeyEnv`, `defaultBackend`, `defaultParseMethod`, `defaultLang`, poll intervals and output limits — are configurable in the DSH GUI settings page or `cordis.patch.yml`.
Is it safe / audited?
The plugin is provided as-is by its maintainer; no independent safety audit is recorded. It is a benign document-parsing integration, not a system-intrusion tool.
07Related DSH workflows
weknora
by tencent
Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.
honcho
by plastic-labs
Memory library for building stateful agents
webnovel-writer
by lingfengqaq
A long-form Chinese web-novel writing assistant for Claude Code and DeepSeek Harness, designed to reduce memory loss and hallucinations in serialized projects of up to 2 million characters.
mirage
by strukto-ai
The World's First Virtual Terminal for AI Agents
08Data and sources
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…
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-09-01. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.