deepseekplugins

DeepSeek Harness plugin

deepseek-harness-file-upload-ocr-plugin

DeepSeek Harness 文件上传与本地 OCR 插件 / File upload and local OCR plugin for DeepSeek Harness: PDF, Word, Excel, PowerPoint, images, and text attachments.

Tools & Capabilities
Jump to install

Source facts

Repository
BYYY-eng/deepseek-harness-file-upload-ocr-plugin
Latest update
Aug 19, 2026
Category
Tools & Capabilities
GitHub stars
4
Format
plugin
Catalog evidence
Upstream dsh.bundle evidence
Evidence path
package.json#dsh.bundle
Checked against
0.1.0-rc.8
Upstream check date
2026-08-20

This evidence comes from the upstream catalog. This site has not installed, run, or security-reviewed the plugin.

Install

Start with a prompt that asks an agent to review the GitHub repository and source. Switch to the command if you want to install it yourself.

Copy this prompt into DSH, Codex, or another agent and ask it to review the GitHub repository and source first.

Do not install or run any commands yet. Read this plugin's GitHub repository, README, and relevant source code. Then answer the questions below clearly and directly so I can decide whether it fits my needs:

1. What is this plugin, and what problem does it solve?
2. Who is it for, and what are its typical use cases?
3. How is it used after installation? Include one minimal example.
4. What known limitations or privacy, security, compatibility, or maintenance risks does it have?
5. Give a clear recommendation: recommend, conditionally recommend, or do not recommend, with reasons.

Distinguish statements documented by the repository, inferences from source code, and unknowns. If evidence is insufficient, say so explicitly. Do not guess or simply repeat the README.

GitHub: https://github.com/BYYY-eng/deepseek-harness-file-upload-ocr-plugin
Plugin: deepseek-harness-file-upload-ocr-plugin
Author: BYYY-eng

Check the source files

Read the README and other files from this plugin directory before installing.

File explorer4 files
README.mdSource · read only
README language

DeepSeek Harness File Upload & OCR Plugin

中文说明 · English

> A file upload and local OCR plugin for DeepSeek Harness. Add PDF, Word, Excel, PowerPoint, image, and text attachments as chat cards; scanned PDFs and images are recognized locally with OCR.

中文项目描述： DeepSeek Harness 文件上传与 OCR 插件：在 Web UI 中以附件卡片上传 PDF、Word、Excel、PowerPoint、图片和文本文件；PDF 与图片在本机自动 OCR 识别。

Features

  • Real attachment cards in the composer and sent conversation messages.
  • Drag files anywhere over the Harness window to upload them.
  • File cards use different accent colors for PDF, images, Word, Excel, PowerPoint, and text files.
  • PDF native-text extraction with automatic OCR for scanned or low-text pages.
  • Local OCR for scanned PDFs and images.
  • DOCX, XLSX/XLSM, PPTX, images, and common text/code formats.
  • Local file processing. Only extracted text is included in the model prompt.
  • Cordis lifecycle, dependency injection, Schemastery configuration, and HMR-compatible registration.

Supported extensions: PDF; PNG, JPEG, WebP, BMP, TIFF; DOCX; XLSX/XLSM; PPTX; TXT, Markdown, CSV, TSV, JSON, XML, YAML, HTML, logs, and common source-code files. Legacy DOC/XLS files must first be saved as DOCX/XLSX.

One-click install

Prerequisites: a working dsh CLI, Python 3.9+, and Git. The installer creates the web profile when it does not exist.

The installer creates the local Python environment first, then registers this package through Harness's official dsh plugin --profile ... add flow. No manual cordis.yml or cordis.patch.yml edits are needed.

Windows PowerShell

git clone https://github.com/BYYY-eng/deepseek-harness-file-upload-ocr-plugin.git
cd deepseek-harness-file-upload-ocr-plugin
powershell -ExecutionPolicy Bypass -File .\scripts\install.ps1 -Profile web
dsh --profile web

Linux / macOS

git clone https://github.com/BYYY-eng/deepseek-harness-file-upload-ocr-plugin.git
cd deepseek-harness-file-upload-ocr-plugin
chmod +x scripts/*.sh
./scripts/install.sh web
dsh --profile web

The first install downloads the Python OCR dependencies. Release branches include prebuilt lib/ files, so pnpm build-script authorization is not required.

GitHub Release archive

If you download a GitHub Release archive instead of cloning the repository, extract it into a folder first, open a terminal in that folder, and run the same platform installer above. This keeps the OCR environment beside the installed bundle. Do not install the .tgz with dsh plugin add alone unless you have also run scripts/setup-ocr.ps1 or scripts/setup-ocr.sh for that installed package.

For the simplest Windows path, choose Source code (zip) on the Release page, extract it, and run scripts\install.ps1.

Official manual bundle install

Advanced users can use the Harness bundle command directly after running the OCR setup script:

powershell -ExecutionPolicy Bypass -File .\scripts\setup-ocr.ps1
dsh plugin --profile web add C:\path\to\deepseek-harness-file-upload-ocr-plugin
dsh --profile web

DeepSeek Harness source checkout

If you run Harness from source and do not have a global dsh command:

powershell -ExecutionPolicy Bypass -File .\scripts\setup-ocr.ps1
cd C:\path\to\deepseek-harness
pnpm dsh plugin --profile web add C:\path\to\deepseek-harness-file-upload-ocr-plugin
pnpm dsh --profile web

Usage

1. Click the document icon in the lower-left corner of the composer. 2. Select one or more files and wait for their cards to appear, or drag files anywhere over the Harness window. 3. Type a question and send it normally.

After sending, the file card remains attached to the message.

Configuration

The bundle provides OCR defaults in cordis.patch.yml. Harness patches replace a row's entire config, so repeat all keys when overriding the file-upload-ocr row.

KeyDefaultMeaning
pythonCommandautoUses .venv created by setup; may be an explicit Python path
maxFileBytes25 MiB (26,214,400 bytes)Maximum size of each uploaded file
maxPages50PDF pages, workbook sheets, or slides
dpi144Scanned-PDF render resolution
nativeTextMinChars24Native PDF characters required to skip OCR
timeoutMs120000Extraction timeout per file
maxOutputChars200000Maximum extracted characters sent to the model

Set DSH_FILE_OCR_PYTHON before startup to override the configured Python runtime without editing YAML.

If Harness reports that the OCR environment is not installed, run the platform setup script again from the plugin folder and restart the web profile.

Privacy and security

Files are processed by a local Python child process. The upload endpoint is same-origin only, rejects path-like filenames, and enforces configured byte, page, time, and output limits. Extracted text is sent to whichever model provider your Harness profile uses; the original file is not uploaded by this plugin.

Open source

Project code is available under the [MIT License](./LICENSE). Dependency licenses and attribution are listed in [THIRD_PARTY_NOTICES.md](./THIRD_PARTY_NOTICES.md). In particular, PDF handling uses permissively licensed pypdfium2/PDFium rather than PyMuPDF/AGPL.

This is an independent community project, not an official DeepSeek release. DeepSeek and DeepSeek Harness are names of their respective owners.

Search terms

DeepSeek Harness upload file plugin · DeepSeek Harness PDF OCR plugin · DeepSeek Harness attachment plugin · DeepSeek Harness 上传文件插件 · DeepSeek Harness 文件附件 OCR

Back to searchOpen GitHub repository

More related plugins

View category