DeepSeek Harness Plugins · Model Inference
Plugins that extend DeepSeek Harness (DSH) with vision, workflows, memory, web tools and more.
Classify animal species from images via URL, Base64, or file path, requiring no extra parameters.
General-purpose speech recognition plugin powered by Whisper Small, supporting multilingual and minority languages via audio URL, Base64, or file path input.
A DSH plugin that detects gestures in an image and returns bounding boxes, confidence scores, and labels, accepting input via URL, Base64, or local file path.
A fast and accurate OCR plugin for DeepSeek Harness that extracts text from images via URL, Base64, or file path, suitable for documents, signage, and screenshots.
An MCP server for MBTI personality testing that guides users through a 28-question simplified or 48-question cognitive-function test and delivers type analysis.
An OCR captcha recognition plugin for DeepSeek Harness that extracts text from common captcha images via file links, Base64 encoding, or local file paths.
Calculates biological age via the Morgan Levine PhenoAge clock from blood biomarkers and provides reference ranges for PhenoAge biomarkers.
A dialect speech recognition plugin for DSH, supporting Mandarin plus 20+ dialects and accents, accepting audio URLs, Base64 data, or local file paths.
An OpenRPC-based MCP plugin that provides JSON-RPC method invocation and discovery services with persistent API key configuration.
Instance segmentation plugin based on YOLO26 that segments 80 COCO object classes, outputting bounding boxes, masks, confidence scores, and labels, with support for URL, BASE64, and file path inputs.
OCR plugin for DeepSeek Harness that extracts permit number, name, gender, date of birth, validity period, and issue location from HK/Macau and Taiwan travel permits, with MRZ machine-readable zone parsing and multiple image input methods.
A DSH human pose estimation plugin powered by YOLO26 that detects people in images and outputs bounding boxes with 17 keypoint coordinates.