Skill UI

Implements standardized function/tool calling across various large language models (LLMs) via OpenRouter. This unified approach allows developers to build sophisticated, multi-step agents and complex workflows, regardless of the underlying provider (e.g., GPT-4o, Claude 3.5). It simplifies the process of achieving structured output and tool-augmented reasoning in production systems.

This guide provides comprehensive instructions for deploying applications powered by the Perplexity Sonar API. It covers setup for major cloud platforms like Vercel, Google Cloud Run, and Docker. Users can deploy production-ready endpoints, manage required secrets, implement edge functions, and utilize caching mechanisms (e.g., Redis) to ensure high availability and performance in production environments.

This guide outlines the comprehensive reference architecture for building advanced, AI-powered search and research applications, modeled after Perplexity. It covers three core tiers: a searchable service with model routing and caching, a citation pipeline for validation and rendering, and an orchestrator for multi-query research synthesis. Use this when designing complex, production-grade, search-augmented LLM integrations or reviewing advanced system patterns.

This skill provides production-ready patterns for integrating with the Perplexity Sonar API in both TypeScript and Python. It covers best practices for creating robust wrappers around the standard OpenAI client, including dedicated handling for citations, search results, and related questions. It also includes essential utilities like implementing exponential backoff for reliable retries. Use this when setting up complex search-augmented generation workflows or standardizing team coding practices around Perplexity usage.

Phoenix is an open-source observability platform for tracing, evaluating, and monitoring LLM applications; use it to debug prompts, run dataset evaluations, build experiment pipelines, and watch production inference in real time.

Fully managed, auto-scaling vector database tailored for production AI workloads such as RAG, recommendations, and semantic search, offering hybrid dense+sparse search, namespaces, metadata filters, and <100ms latency on serverless infrastructure.

Prediction Monitor Assistant automates machine learning deployment monitoring tasks, offering step-by-step guidance, best-practice patterns, production-ready code and configuration generation, and validation to keep serving and inference workflows reliable.

Automates PyTorch training workflows by guiding data preparation, model configuration, hyperparameter tuning, and experiment tracking while following industry best practices for production readiness.

Qdrant is a high-performance, Rust-powered vector database designed for production-grade RAG (Retrieval-Augmented Generation) and semantic search applications. It provides fast nearest neighbor search, supports complex hybrid search using metadata filtering, and offers horizontal scalability through sharding and replication. Ideal for building real-time recommendation systems and enterprise-level knowledge retrieval.

Designs and implements production-grade RAG systems with document chunking, embeddings, vector stores, hybrid search, reranking, and evaluation, ideal for knowledge-grounded AI apps that need semantic retrieval, similarity search, and context augmentation.

This guide provides a comprehensive production checklist and SDK pattern for integrating Retell AI into voice agents and telephony systems. It covers the necessary steps, error handling, and implementation patterns for deploying reliable, AI-powered phone call automation and voice interactions in a production environment.

This resource provides production-ready, structured patterns for building robust voice agent applications using the Retell AI SDK. It covers best practices such as implementing singleton client management, defining typed agent configurations with integrated LLMs, adding automatic retry logic for call failures, and managing large-scale outbound campaigns. Use this when building stable, scalable, and reliable voice automation systems.