A simple example showing how to do metadata extraction with RAG using Mastra, OpenAI, and PGVector.
Music metadata parser for Node.js, supporting virtual any audio and tag format.
Create and modify PDF files with JavaScript
various machine learning routines for node
tar for node
Lightweight dependency injection container for JavaScript/TypeScript
DOM to Semantic-Markdown for use in LLMs
A lightweight font data library — fontkit’s little sibling
GeoTIFF image decoding in JavaScript
Read text and parse tables from PDF files. Supports tabular data with automatic column detection, and rule-based parsing.
Tokenized zip support
High performance Node.js image processing, the fastest module to resize JPEG, PNG, WebP, GIF, AVIF and TIFF images
Unofficial Zalo API for JavaScript
Extract clean, timestamped YouTube captions, subtitles, transcripts, and video metadata for AI summaries, RAG, search, and slide-ready workflows.
official twitter text linkification
Get the metadata from a Google Cloud Platform environment
Polyfill for Metadata Reflection API
Light ECMAScript (JavaScript) Value Notation - human written, concise, typed, flexible
No description provided.
Build swagger-compliant REST APIs using TypeScript and Node
Language metadata and dynamic loading for the CodeMirror code editor
A unified, high-performance i18next CLI.
JavaScript version of GPAC's MP4Box tool
remark plugin to support frontmatter (yaml, toml, and more)
Provides apis for extracting common metadata out of files as well as low level apis for advanced metadata parsing. Currently exif (jpeg/jpg) is almost entirely supported and mpeg4 (mp4,m4v,moov...) has limited support. For common metadata the FileInfo class provides methods names after the metadata items taking a filename. As an example, to get the origin date of a file you would call FileData::FileInfo.origin_date(filename). Advanced apis are provided via specific classes for each metadata type. For example, Exif for exif data and Mpeg4 for mpeg4 data. These can be used to improve the performance of gathering multiple metadata values from a file
Command-line tool that automatises photo/video uploads to Flickr. Entering 'flickru <directory>' in your command line, any photos under 'directory' (and subdirs) are uploaded to your Flickr account (interactively entered the first time you start flickru). Photos are identified by case-insensitive extensions: GIF, JPEG, JPG, PNG, and TIFF. Videos are identified by case-insensitive extensions: AVI, MPEG, and MPG. flickru automatically sets the following Flickr metadata: (1) date taken: file last-modification time, unless JPEG/TIFF Exif metadatum 'date_time_original' is found (Flickr understands it natively). (2) privacy policy: private, visible by friends & family, hidden for public searches (3) safety level: safe (4) permissions: friends & family can add comments to the photo and its notes; nobody can add notes and tags to the photo (5) description: for videos longer than 90s (Flickr's longest allowed duration) but shorter than 500MB (Flickr's maximum permisible size), it will contain an annotation about its large duration. (6) title: extracted from the parent directory name (7) geolocation & accuracy: extracted from the parent directory name, unless JPEG/TIFF Exif GPS metadata is found (Flickr understands them natively). Before uploading photos, please, make sure that you have correctly named each photos parent directory according to the name format 'TITLE[@LOCATION[#PRECISION]]', where: (1) TITLE is the desired title for the photos stored in the directory. If no LOCATION is given, flickru tries to extract the location from Wikipedia page TITLE. (2) LOCATION is the location of the photos, specified as: (a) the Wikipedia page name (whitespaces allowed) of the location or (b) its coordinates LATITUDE,LONGITUDE (3) PRECISION is the Flickr geolocation precision. Flickru sets it to one of the following case insentitive literals: 'street', 'city', 'region', 'country', 'world'. Photos are classified into photosets. If the photoset does not exist, flickru creates it. This photoset is named after its grandparent directory. The photoset is arranged by 'date taken' (older first). To see some examples on the directory structure recognised by flickru, please explore the subdirectories under 'var/ts'. GitHub : http://github.com/jesuspv/flickru RubyGems: http://rubygems.org/gems/flickru