MinerU CLI
MinerU is a CLI with a document library: search and revisit an already parsed report using its local cache.
- 83,203 ★ on the repository
- Last commit
Extended access
Why this level
- The CLI reads documents and saves a library cache; remote mode sends a document to an external parser.
- Doclib collects diagnostic telemetry by default, excluding document contents.
Install
Manual install
npx skills add opendatalab/MinerU --skill mineru --global --yesFirst README agent method: installs the skill, not the CLI.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The CLI reads documents and saves a library cache; remote mode sends a document to an external parser.
- Doclib collects diagnostic telemetry by default, excluding document contents.
This is third-party code. Review the repository files before installing.
What it does
MinerU is a CLI for parsing PDFs, images and office documents into structured text. Users choose a processing tier, receive results and continue at a specific page or block. A local document library caches parsed material for repeated search and reading. The separate `mineru-kit parse` command performs stateless conversion without the library. An agent skill is documented separately from CLI installation. Documents are parsed locally by default; remote parsing requires an explicit choice.
Who it is for. Long-report analysts finding passages, revisiting pages and retaining source locators.
Good fit when
- An analyst reads a long PDF in several bounded parts.
- An agent searches a local library and opens the matched page.
- A team converts an office document to Markdown for analysis.
- An operator selects a scan-processing tier for available hardware.
Not a fit when
- A user installs only the skill and expects a CLI executable without separate setup.
- A confidential document would be sent to a remote parser without approval.
- A team treats Flash as final reading quality for a complex PDF without comparison.
Example request
Parse a local report copy with MinerU, find expense references and show source-page locators without remote upload.Limitations
The current branch requires Python 3.11–3.14, so older environments need replacement before upgrading. Office formats use local Flash parsing, while PDFs and images support all four quality tiers. Neural tiers need model downloads and sufficient memory, with requirements determined by the chosen engine. Doclib telemetry starts enabled and can be disabled with `mineru telemetry disable`; this telemetry scope excludes standalone `mineru-kit parse`.
How to disable. Finish jobs and run `mineru server stop` if using the library server. Stop a standalone converter separately; document caches and exports remain.
FAQ
How do skill and CLI installation differ?
`npx skills add opendatalab/MinerU --skill mineru --global --yes` adds document-handling instructions to the agent. The CLI executable needs a separate Python-environment installation; follow the manual setup in the next answer.
How does manual CLI setup begin?
The author starts with `pip install uv`, then `uv venv .mineru --python 3.12`. Activate that environment for your OS and run `uv pip install -U "mineru>=4.0,<5"`.
How do I create Markdown without the library?
Quick Start gives `mineru-kit parse document.pdf -o document.md --tier standard`. Substitute the source and a new output filename, then inspect tables and text order.
What does server_not_running require?
Run `mineru server start` for the library service. `mineru server status --json` then shows its state and available tiers; choose a supported local tier before parsing again.
How do I continue reading a large document?
After reading part of a document, the result includes `next_request.locator`, a pointer to the next section. Ask the agent to continue with `mineru read` using that pointer to retrieve the next part with its page references.
Discussion
No questions or reviews yet.
Related
CLI toolsA zero-key, zero-dependency skill and CLI that turns PDF, Office and image files into Markdown via MinerU's cloud.
A local CLI extracts text from documents and scans, searches archives and prepares source material for an agent’s answer.
Docling is a CLI for mixed document archives, bringing PDFs, office files and spreadsheets into a shared format for AI search.
A PDF-reading MCP server: structured text, tables, OCR, visual crops and page-level citations
pdfcpu validates, merges and splits PDFs, extracts content and manages pages through a dedicated command-line interface.
Two skills for Claude Code and Codex: the agent translates papers in the EasyRead library and answers your margin notes
README in short
The README distinguishes the library CLI, standalone converter, API and WebUI. Four tiers have different model and resource requirements. Code terms are based on Apache 2.0 with additional conditions in LICENSE.md.