Skill UI

A comprehensive guide to optimizing LLM API usage within LangChain frameworks. Techniques include implementing token usage tracking, model tiering to route tasks based on complexity, utilizing caching to eliminate duplicate calls, prompt compression for efficiency, and enforcing strict budget limits to manage operational costs effectively.

This guide provides advanced techniques for managing API rate limits (429 errors) and complexity budgets when interacting with the Linear API. It covers reading rate limit headers, implementing exponential backoff with jitter for robust retries, controlling request bursts via token buckets, and optimizing GraphQL queries to minimize complexity scores. Use this when developing reliable integrations that must adhere to strict API quotas.

This toolkit provides a comprehensive framework for managing and optimizing API costs when using Mistral AI. It includes functions for calculating token-based costs across various models, implementing a smart model router based on task complexity, and managing daily/monthly budgets with critical alerts. Furthermore, it offers methods for prompt optimization to reduce token consumption, making it essential for large-scale application cost control and billing analysis.

This skill provides a comprehensive guide to understanding OpenRouter's API billing structure. It teaches users how to accurately calculate costs based on prompt (input) and completion (output) tokens. Users can utilize provided scripts to estimate costs, check credit balances, and implement cost-saving strategies like using `:floor` or `:free` model variants. Ideal for budgeting, performance comparison, and production cost tracking.

This skill defines canonical, ordered middleware patterns for LangChain and LangGraph agents. It addresses critical production-level LLM deployment issues, including PII leakage across tenants, token budget overconsumption, prompt injection defense (Guardrails), and cache key corruption. Use this to build secure, scalable, and cost-aware AI applications.