Skill UI

A comprehensive guide on implementing failover and fallback mechanisms for Large Language Models (LLMs) when using OpenRouter. This skill covers native model fallbacks, provider routing strategies, and advanced client-side fallback chains. Use this pattern when building mission-critical or highly resilient systems that must maintain service availability even if a primary model or an entire provider goes offline. It ensures robust API integration and graceful degradation.

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 a comprehensive quickstart for integrating with OpenRouter, a unified endpoint for accessing multiple large language models (LLMs). It includes examples in cURL, Python, and TypeScript, demonstrating how to send chat completion requests, handle complex JSON responses, manage model parameters, and calculate usage cost across various providers (e.g., Llama, GPT-4o, Claude). Ideal for developers testing API connectivity or building multi-model applications.

This guide provides a comprehensive walkthrough for setting up and authenticating with the OpenRouter API. It covers generating secure API keys, configuring environment variables, and initializing clients in Python and TypeScript. Learn best practices for secure integration, including key management, verification, and key rotation, ensuring reliable connectivity to various large language models (LLMs).

This skill provides comprehensive tools to monitor the real-time availability, status, and performance of various large language models (LLMs) accessed via OpenRouter. It supports active health probes, catalog checks, and scheduled monitoring via Python and Bash, ensuring robust fault detection and failover logic for production systems that rely on external AI APIs.

This skill provides comprehensive best practices and Python code examples for optimizing API calls to Large Language Models (LLMs) via OpenRouter. It covers benchmarking model latency, implementing streaming to achieve low Time-To-First-Token (TTFT), and processing requests concurrently using asynchronous programming (asyncio). Use this when developing real-time, high-throughput applications that require predictable and fast model responses.

Optimizes prompts for LLMs to cut token usage, lower costs, and boost response quality by analyzing, condensing, and rewriting instructions; useful when teams need faster, cheaper, clearer AI conversations.

Covers prompt and response caching plus Cache Augmented Generation patterns for LLMs such as Claude and OpenAI to reuse stable context, reduce repeat work, and cut latency/cost on recurring queries.

This skill specializes in designing, optimizing, and evaluating prompts for Large Language Models (LLMs). It allows users to generate optimized prompt templates, structured output schemas (e.g., JSON mode, function calling), and comprehensive test suites. Use this skill when building new LLM applications, implementing complex reasoning patterns like Chain-of-Thought, defining system guardrails, or building robust evaluation frameworks to measure and improve model performance.

Meta’s Prompt Guard is an 86M-parameter classifier that spots prompt injections and jailbreak attempts before they reach LLMs, filtering user input, third-party data, or RAG documents with multilingual support and sub-2ms latency for fast deployments.

A comprehensive collection of battle-tested, structured prompts designed to elevate interactions with Large Language Models (LLMs). Covers advanced use cases including expert coding reviews, system architecture design, performance analysis, technical writing, and creative brainstorming, ensuring high-quality, actionable AI output.

Quantizes HuggingFace LLMs to 8-bit or 4-bit with bitsandbytes, cutting memory by up to 75% while keeping accuracy, and supports QLoRA fine-tuning plus 8-bit optimizers for faster, memory-efficient training.