Packages

Serverless AI inference platform running open-weight models at Cloudflare's global edge network. Supports text generation (Llama, Mistral), embeddings, image classification, speech-to-text, and translation via REST API — no GPU setup required.

Backend development platform that combines a type-safe backend framework with automatic infrastructure provisioning. Encore parses TypeScript/Go source code annotations (API endpoints, services, databases, queues) and automatically provisions the required cloud infrastructure on AWS, GCP, or Azure — or runs a complete local development environment with all infrastructure. Generates OpenAPI docs, distributed tracing, and architecture diagrams automatically from code.

In-memory distributed data grid and stream processing platform. Hazelcast embeds in Java applications or runs as a standalone cluster. Core data structures: IMap (distributed HashMap with TTL, near-cache, entry listeners), IQueue, ISet, IList, ITopic (pub/sub), AtomicLong, Lock, Semaphore — all distributed across cluster members. Hazelcast Jet (unified in 5.x) enables distributed stream processing with SQL over streams. Spring Cache integration via @Cacheable maps to IMap. Hazelcast Platform includes IMDG + Jet. Use cases: session clustering, distributed locking, agent result caching across multiple service instances, and distributed counters.

MCP server for Jimeng AI (即梦) — ByteDance's generative AI platform for creating images and videos from prompts. Enables AI agents to generate images and videos, mix images, use reference images, and create visual content programmatically through ByteDance's Jimeng AI service.

Managed Farcaster API that provides a clean REST interface to Farcaster social data without running your own Hub node. Neynar handles the Hub infrastructure and provides endpoints for user lookup, cast feeds, reactions, follows, channel data, Frames interactions, and posting casts (including managed Signer creation). The de facto standard way to build Farcaster applications via API.

Reactive web framework for Scala and Java built on Akka Streams and Netty. Play is the leading full-stack Scala web framework — handles HTTP routing, JSON/HTML templating, WebSockets, server-sent events, and reactive database access. Play 3.x (2023) migrated from Akka to Apache Pekko (Akka's open-source fork) due to Akka's license change. Provides non-blocking async request handling via Pekko Streams, a routing DSL for Scala and Java, built-in JSON support via play-json, and Twirl template engine for HTML. Used by LinkedIn, The Guardian, and major financial institutions.

PostGrid REST API — programmatic physical mail sending (letters, postcards, checks) and address verification, enabling agents to automate direct mail campaigns, compliance notices, and physical document delivery without manual print-and-mail workflows.

Cross-platform GUI automation library — controls mouse, keyboard, and takes screenshots programmatically. PyAutoGUI features: pyautogui.moveTo(x, y), pyautogui.click(), pyautogui.typewrite('text'), pyautogui.press('enter'), pyautogui.screenshot(), pyautogui.locateOnScreen() (image template matching), pyautogui.hotkey('ctrl', 'c'), pyautogui.scroll(), drag and drop, window management, PAUSE between actions, FAILSAFE (move mouse to corner to abort), and cross-platform support (Windows, macOS, Linux). Primary Python library for desktop GUI automation for agent computer-use tasks.

Async ORM for Rust built on SQLx with a familiar Active Record and Data Mapper pattern. Provides entity definitions, query building, migrations (SeaMigration), and model relations with async execution. Positioned as the Rust equivalent of ActiveRecord or Hibernate — generate entities from existing databases or define code-first. Part of the SeaQL ecosystem (SeaQuery for query builder, SeaMigration for migrations).

Deno-based serverless edge functions tightly integrated with Supabase's Postgres database, auth, and storage, deployable via REST API or CLI for backend logic close to users.

TypeScript toolkit for building AI-powered web applications with first-class streaming support, React/Next.js hooks, multi-provider abstraction, and tool/agent capabilities.

Frontend cloud platform with git-based deployments, edge network, serverless functions, and native Next.js integration for deploying web apps globally.

AWS DynamoDB is a fully managed NoSQL key-value and document database with single-digit millisecond latency at any scale — accessed via AWS SDK or DynamoDB REST API with IAM auth.

Cloud-native platform that serves as a centralized MCP registry, gateway, and orchestrator with enterprise governance, security controls, cost monitoring, and prompt injection defense for managing MCP servers across organizations.

Apple's reactive programming framework for Swift. Provides Publisher/Subscriber model for composing async operations — similar to RxSwift but built into the platform. Combine is integrated with SwiftUI (@Published, ObservableObject), URLSession, NotificationCenter, and Core Data. Use operators like map, filter, merge, combineLatest to compose async data streams from agent networking, user input, and system events.

The most widely-used authentication library for Ruby on Rails. Devise provides a modular authentication solution with 10 configurable modules: database_authenticatable (password hashing), registerable (user signup), recoverable (password reset), rememberable (remember me cookies), trackable (login tracking), validatable (email/password validation), confirmable (email confirmation), lockable, timeoutable, and omniauthable (OAuth via OmniAuth). Generates migrations, routes, views, and controllers.

Computer vision dataset management and model evaluation platform — explore, evaluate, and curate CV datasets. FiftyOne features: fo.Dataset for loading images/videos with labels, fo.launch_app() interactive web UI (tags, filtering, similarity search), model evaluation (evaluate_detections, evaluate_classifications), dataset zoo (fo.zoo.load_zoo_dataset), label studio integration, COCO/YOLO/VOC import/export, brain plugins (compute_similarity for duplicate finding, compute_uniqueness, compute_hardness), active learning workflows, find_duplicates(), and dataset slices. Used for pre-training data curation, model failure analysis, and active learning for agent computer vision.

HTTP request interceptor library for Python testing. HTTPretty patches Python's socket library to intercept all HTTP connections — it doesn't require changing how HTTP clients are created. Supports all major Python HTTP libraries (requests, httpx, urllib3, boto3, aiohttp) and allows registering fake responses for specific URLs. Used for unit testing code that makes HTTP requests without mocking individual client instances.

MCP server that bridges Home Assistant smart home instances with LLMs, enabling natural language control and monitoring of smart home devices. Supports lights, climate, covers, switches, sensors, media players, fans, locks, vacuums, cameras, add-on management via Supervisor, HACS package management, automation creation/editing, and real-time SSE state updates.

Modern project management API for software teams covering issues, projects, cycles, roadmaps, and team workflows via a rich GraphQL interface.