Packages

PostgreSQL extension for time-series data that adds automatic partitioning (hypertables), continuous aggregations, and compression while preserving full SQL compatibility and the PostgreSQL ecosystem.

Serverless Kafka-compatible message streaming service with a REST API optimized for edge and serverless environments — pay-per-message pricing with zero cluster management and HTTP-based produce/consume for environments where native Kafka TCP connections are not available.

High-performance reactive database for React Native and web apps. WatermelonDB uses SQLite (native) or IndexedDB (web) for local storage with a reactive observation system that automatically re-renders components when data changes. Designed for 'lazy loading' — queries run lazily and only fetch data when observed. The sync engine enables server synchronization with a push/pull protocol that your server implements. Used by Nozbe (productivity apps) in production with tens of thousands of records.

A standardized framework for recording browser user interactions via a Chrome extension and replaying them either exactly or as AI-condensed smart sequences, with the ability to convert recordings into reusable MCP servers for AI agent consumption.

Serverless PostgreSQL-compatible database with built-in full-text search powered by Elasticsearch and native AI/vector search features, accessible via REST API and TypeScript SDK.

Web indexing and search tool that crawls websites, generates vector embeddings, and exposes crawled content through an MCP server and REST API for LLM agents.

Official ElevenLabs MCP server enabling AI assistants to generate speech, clone voices, transcribe audio, isolate voices, and create soundscapes through the ElevenLabs API.

MCP server that enables AI assistants to control FreeCAD for parametric 3D modeling, including creating/editing objects, converting 2D drawings to 3D, executing arbitrary Python in FreeCAD, and accessing the parts library.

Git-like versioning layer for data lakes built on object storage (S3, GCS, Azure Blob). lakeFS adds branches, commits, and merges to data stored in object storage — enabling data teams to create isolated development environments, test pipeline changes, and roll back bad data transformations. Works transparently with existing tools (Spark, Presto, dbt, Airflow) via an S3-compatible API. No data movement — versioning is metadata-only.

Python bindings for YARA — pattern matching framework for malware identification and threat hunting. yara-python features: yara.compile() from string or file, rules.match() on files/bytes/processes, YARA rule syntax (strings: patterns, conditions: logic), string pattern types (text, hex, regex, base64), meta for rule metadata, tags for categorization, yara.Rules object, match.strings for matched locations, process scanning (pid=PID), externals for context variables, private rules, and scan timeout control. Standard malware analysis and threat hunting tool — used by EDR vendors, CSIRT teams, and malware researchers for file classification.

Commission-free US equity and crypto brokerage API with paper trading, real-time market data streaming, and broker API for building trading products on top of Alpaca's regulated clearing infrastructure.

Bentley iTwin Platform REST API for infrastructure digital twin creation and management platform. Enables AI agents to manage iModel (infrastructure model) creation and synchronization from design files, handle change management and design review workflows across disciplines, access real-time IoT sensor data integration with infrastructure models, retrieve 3D visualization and clash detection results for engineering review, manage project and team collaboration on large infrastructure programs, handle CDE (Common Data Environment) document and version management, access infrastructure asset lifecycle data from design through operations, retrieve GIS and geospatial coordinate integration for infrastructure siting, manage reality capture and point cloud integration workflows, and integrate infrastructure digital twin data with project management, GIS, and operations platforms.

Cloud-based image and video management platform with a transformation pipeline, AI-powered tagging, and CDN delivery accessible via REST API and SDKs.

Task-oriented AI orchestration framework by Prefect where you define Tasks with explicit goals and completion criteria, and AI agents collaborate to complete them.

Provides WebRTC video and audio infrastructure via REST API and client SDK, enabling agents to create rooms, issue access tokens, trigger cloud recording, and receive real-time transcription without managing media servers.

Microsoft's distributed training and inference optimization library for large-scale deep learning. DeepSpeed enables training LLMs with billions of parameters on GPU clusters via ZeRO (Zero Redundancy Optimizer) — which shards optimizer state, gradients, and model parameters across GPUs to dramatically reduce memory footprint. Also provides inference optimization (DeepSpeed-Inference, DeepSpeed-MII) with kernel fusion, quantization, and tensor parallelism for faster LLM serving. Used to train GPT-3, Megatron-DeepSpeed, and many other large models.

Dyte REST API — highly customizable real-time video and audio meeting infrastructure with programmable UI components, enabling agents to create meetings, manage participants, control recording, and build branded video experiences.

Minimalist, high-performance Python web framework for building REST APIs and microservices. Faster than Flask and Django for pure API use cases — designed for production API workloads with minimal overhead. Supports WSGI and ASGI. Resource-based routing with explicit request/response objects. No view templates, forms, or ORM — pure API framework.

Express.js-inspired Go web framework built on fasthttp instead of the standard net/http. Fiber prioritizes developer experience from Node.js backgrounds with familiar API patterns (app.Get, app.Post, c.JSON, c.Params) while delivering near-zero-memory-allocation performance. Ideal for building REST APIs and microservices where Go developers want Express ergonomics with Go performance.

Constrains LLM generation at the token level using regex, JSON schemas, and context-free grammars, producing structured output faster and more reliably than prompt-only approaches.