Packages

Self-hosted AI coding assistant with a REST API. Tabby runs locally or on-premises, providing GitHub Copilot-like code completions using open-source LLMs (CodeLlama, StarCoder, DeepSeek Coder). Ships with IDE plugins (VS Code, JetBrains, Vim/Neovim) and a REST API for custom integrations. Enterprise features include RBAC, Git context indexing, and a team management UI. Core value: full data privacy with no code sent to external services.

The most popular server-side Swift web framework. Vapor builds on Apple's SwiftNIO for high-performance async networking. Includes routing, middleware, Fluent ORM (PostgreSQL, MySQL, SQLite), Leaf templating, WebSocket support, and async/await throughout. Used for building backend APIs that share Swift code with iOS/macOS clients, enabling full-stack Swift development.

Async Python AMQP client for RabbitMQ built on asyncio. aio-pika is the asyncio-native wrapper around pamqp that provides a clean API for publishing and consuming RabbitMQ messages in async Python applications. Supports all AMQP patterns: direct exchange, topic exchange, fanout, dead letter queues, and RPC patterns. The go-to async RabbitMQ client for Python.

Fast approximate nearest neighbor (ANN) search library implementing HNSW (Hierarchical Navigable Small World) algorithm. hnswlib features: hnswlib.Index for in-memory vector index, add_items() for batch insertion, knn_query() for k-nearest-neighbor search, save_index()/load_index() for persistence, ef_construction and M parameters for accuracy/speed trade-off, ef (query-time) parameter, multi-threading for batch queries, mark_deleted() for soft deletion, and both L2 (Euclidean) and cosine space metrics. 10-100x faster than brute-force search at slight accuracy cost. Lightweight alternative to FAISS for CPU-only environments.

Minimal, purely functional HTTP library for Scala that provides both server (http4s-ember-server, http4s-blaze-server) and client (http4s-ember-client) implementations. HTTP requests and responses are modeled as pure data types with fs2 streaming bodies. Routes are PartialFunction[Request[F], F[Response[F]]] composable values — no reflection, no annotations. Part of the Typelevel ecosystem alongside Cats Effect and Doobie.

MCP proxy server written in Go that aggregates multiple MCP resource servers behind a single HTTP endpoint. Supports SSE and streamable HTTP transport modes, connects to upstream servers via stdio, SSE, or streamable-HTTP, and consolidates tools, prompts, and resources from all backends.

Python library for controlling and automating interactive terminal applications — spawns child processes and interacts with them via pseudo-terminal (pty), similar to Unix expect. pexpect features: pexpect.spawn() for process spawning, expect() with string/regex/EOF/TIMEOUT patterns, sendline()/send() for input, before/after attributes for matched text, interact() for handoff to user, pexpect.run() for simple one-shot execution, pxssh class for SSH automation, timeout and maxread configuration, logfile for interaction logging, and fdpexpect for existing file descriptors. Essential for automating CLI tools that don't support piped stdin/stdout.

Amazon Simple Email Service (SES) — high-volume transactional and bulk email sending via AWS with SMTP and REST API interfaces, pay-per-use pricing starting at $0.10/1000 emails, and deep AWS ecosystem integration.

Runs Deno (TypeScript-native) scripts globally at the edge using V8 isolates, with built-in Deno KV for persistent storage and zero-config deployment from GitHub or the Deno CLI.

EdgeDB REST API enables agents to query and mutate a graph-relational database using EdgeQL or SQL over HTTP, with strong type safety, schema migrations, and automatic link traversal replacing manual joins.

KOMOJU Japan multi-payment gateway REST API for international merchants entering Japan to accept credit cards, convenience store (konbini) payments, bank transfers, PayPay, PayEasy, and other Japan-specific payment methods through a single unified integration without requiring Japanese corporate entity. Enables AI agents to manage KOMOJU payment session creation for Japan multi-method checkout automation, handle credit card processing for Japan card payment automation, access konbini convenience store payment for Japan cash payment automation, retrieve bank transfer (Furikomi) for Japan bank payment automation, manage PayPay and LINE Pay for Japan mobile wallet automation, handle payment status polling for Japan async payment confirmation automation, access refund and cancellation for Japan order management automation, retrieve webhook notification for Japan payment event automation, manage subscription billing for Japan recurring payment automation, and integrate KOMOJU with international merchants, marketplaces, and games companies for end-to-end Japan payment acceptance automation.

Real-time application performance monitoring dashboard for Laravel — tracks server metrics, slow queries, slow jobs, exceptions, HTTP requests, and cache hits with minimal configuration. Pulse features: /pulse dashboard with configurable cards, server monitoring (CPU, memory, disk via pulse:check command), slow query detection (DB::listen threshold), failed/slow job tracking, exception aggregation, cache interaction stats, user activity timeline, custom recorders, and database-backed metric storage. Ships as laravel/pulse package with Livewire dashboard component. Requires no external services — all data stored in your app's database (pulse_* tables). Complements Laravel Telescope for development with production-grade metric aggregation.

Matterport REST and GraphQL API for 3D digital twin and spatial data platform. Enables AI agents to manage 3D space creation and scan processing workflows, handle virtual tour embedding and sharing automation, access spatial data and floor plan extraction, retrieve mattertag and annotation data from 3D spaces, manage space organization and folder hierarchy, handle model processing status and completion tracking, access point cloud and mesh data for construction and AEC workflows, retrieve room and area measurement data, manage team collaboration and access permissions, and integrate 3D spatial data with real estate, construction, insurance, and facility management platforms.

Classic Python NLP toolkit — comprehensive library for text processing, linguistic analysis, and NLP tasks. NLTK features: tokenization (word_tokenize, sent_tokenize), POS tagging (pos_tag), named entity recognition (ne_chunk), stemming (PorterStemmer, WordNetLemmatizer), sentiment analysis (VADER SentimentIntensityAnalyzer), stopwords corpus, WordNet integration, frequency distributions (FreqDist), chunking, parsing, and 100+ corpora and language resources. nltk.download() for corpus installation. Pre-transformer NLP for agent text preprocessing and feature extraction pipelines.

Provides a unified AI gateway API for routing requests across multiple LLM providers (OpenAI, Anthropic, Gemini, etc.) with built-in observability, semantic caching, rate limiting, fallback strategies, and prompt management.

Python-native workflow orchestration as a service — wrap any Python function with @flow and @task decorators to get managed scheduling, retries, observability, and a REST API for triggering and monitoring runs.

Production model serving library built on Ray for scalable ML inference. Ray Serve supports model composition (chaining models in DAGs), dynamic batching, autoscaling, and mixed hardware deployments. REST API automatically exposed for deployed models. Anyscale provides managed Ray/Ray Serve. Supports any ML framework (PyTorch, TensorFlow, HuggingFace, vLLM) and handles LLM serving natively.

Enterprise CRM platform API for accessing and manipulating Salesforce data including contacts, opportunities, accounts, cases, and custom objects through REST, SOAP, Bulk, and Streaming APIs.

Python concurrency library using greenlets (cooperative coroutines) with monkey-patching to make blocking stdlib calls non-blocking. gevent.monkey.patch_all() transparently makes socket, threading, and I/O operations async without code changes — existing synchronous code becomes concurrent. Powers Gunicorn's gevent worker and used in legacy Python services for I/O concurrency without async/await syntax.

A single-binary Go AI bot platform that combines MCP server capabilities, 19+ messaging platform integrations, 75+ built-in tools (file ops, shell, browser automation, calendar), and support for 16 AI backends into one zero-dependency executable.