Where is AI automation heading?

An AI pair programmer that runs in your terminal and edits code directly in your repo. Developers use it to write features, fix bugs, and refactor code through natural language — replacing some of the work junior developers do.

Connects Vercel's AI SDK to Anthropic Claude models. Growth here reflects Claude adoption specifically within production web applications.

Lets AI agents control a web browser — clicking buttons, filling forms, navigating pages. Used to automate repetitive web-based tasks that currently require a human at a keyboard.

Also in: Office & Administrative

The official SDK for Model Context Protocol — an emerging open standard that lets AI models connect to external tools, databases, and APIs. Its explosive growth signals that AI tool integration is being standardized.

A framework for orchestrating multiple AI agents that collaborate on complex tasks. Each agent has a role (researcher, writer, reviewer) and they pass work between each other — mimicking how human teams operate.

Also in: Office & Administrative

Connects Vercel's AI SDK to OpenAI models. When developers build AI features into web apps using the AI SDK, this is one of the most common provider packages they install.

Vercel's AI SDK — the most popular way to add streaming AI chat, text generation, and tool use to React and Next.js web apps. Its growth tracks how fast AI is being embedded into production web products.

Also in: Office & Administrative

Hugging Face's lightweight agent framework. Designed for quick automation tasks where a full multi-agent system is overkill — one AI agent that can call tools and take actions.

Microsoft's framework for building multi-agent AI systems where agents can have conversations, write code, and coordinate to solve complex problems without human intervention.

Alternate distribution of Microsoft AutoGen. Same multi-agent capabilities — tracked separately because both package names see significant download volume.

IBM's document conversion library that turns PDFs, Word docs, and scanned images into structured data an AI can analyze. Heavily used in legal tech for processing contracts, court filings, and regulatory documents.

JavaScript SDK for LlamaIndex — builds AI systems that can search, summarize, and answer questions across large document collections. Used in legal tech for contract review and research automation.

Converts PDFs into clean Markdown text while preserving structure like headings, tables, and lists. Critical for legal and compliance workflows where document formatting carries meaning.

The JavaScript core of LangChain — provides chains, agents, and retrieval tools for building AI document processing pipelines in web applications. A cross-industry building block for AI-powered workflows.

Also in: Finance & Insurance, Office & Administrative

An ETL pipeline for documents — ingests PDFs, emails, HTML, images, and Office files, then outputs clean structured data for AI processing. Used across industries wherever large document backlogs need to be made AI-readable.

Also in: Finance & Insurance, Office & Administrative

The standard Python library for reading and manipulating PDF files. A foundational dependency in any workflow that processes contracts, financial statements, or regulatory filings at scale.

Also in: Finance & Insurance

Lets users query spreadsheets and databases using plain English instead of writing code. An analyst can ask 'show me Q3 revenue by region' and get the answer — automating work that previously required SQL or Excel expertise.

Connects LangChain's JavaScript framework to OpenAI models. Used in production web apps that need GPT-4 for document analysis, data extraction, and automated decision-making workflows.

Also in: Office & Administrative

An open-source vector database that stores document embeddings for AI-powered search. When financial firms build AI systems to search across thousands of research reports or filings, ChromaDB is a common backend choice.

Python framework for building 'retrieval-augmented generation' (RAG) systems — AI that answers questions by searching through your own documents first. Used in finance for research, compliance, and due diligence automation.

Also in: Office & Administrative

Automates the tedious work of cleaning, validating, and preparing datasets for analysis. Handles missing values, type conversions, and formatting — tasks that data analysts and junior quants spend hours on manually.

NLP models trained on biomedical text — can identify drug names, gene references, and disease mentions in research papers and clinical documents. Used in pharmaceutical research and clinical data extraction.

A clinical NLP toolkit built on spaCy that extracts structured data from unstructured medical records — identifying diagnoses, medications, and procedures from doctor's notes and discharge summaries.

JavaScript client for Replicate's AI model platform. Used by web developers to add AI image generation, video processing, and audio tools directly into websites and web applications.

An optimized implementation of OpenAI's Whisper speech-to-text model that runs 4x faster. Used to transcribe podcasts, meetings, video content, and interviews — automating work done by human transcriptionists.

OpenAI's original Whisper model for speech recognition and transcription. Supports 99 languages and is used for subtitling, podcast transcription, voice-to-text features, and accessibility tools.

Stability AI's official SDK for their image generation models (Stable Diffusion). Used by developers building AI-powered design tools, marketing asset generators, and creative automation pipelines.

A cloud platform for running AI models via API — especially popular for image generation, video creation, and audio processing. Developers use it when they want AI creative tools without managing GPU infrastructure.

Hugging Face's library for running Stable Diffusion, FLUX, and other AI image generation models. The main open-source tool that lets developers generate, edit, and transform images with AI — widely used in design and media production.