LLM Guides

These guides cover the most important patterns for working with language models in SynapseKit. Each guide is self-contained and ends with a complete working example you can run immediately.

Guides in this section

Guide	What you'll build	Difficulty	Time
LLM Provider Comparison	Side-by-side benchmark across OpenAI, Anthropic, Groq, and Ollama	Beginner	~15 min
Cost-Aware LLM Router	Complexity classifier + routing table + circuit breaker + budget guard	Intermediate	~20 min
LLM Fallback Chains	Primary → secondary → tertiary failover with CircuitBreaker	Intermediate	~15 min
Semantic Response Caching	SQLite and Redis cache backends, cache hit/miss metrics	Beginner	~15 min
Structured Output with Pydantic	Pydantic BaseModel response schemas, JSON mode, field validation	Beginner	~15 min

Prerequisites

All guides assume you have SynapseKit installed:

pip install synapsekit

Individual guides list any additional provider extras (e.g. synapsekit[openai,groq]) in their Prerequisites section.

Which guide should I start with?

• New to SynapseKit? Start with LLM Provider Comparison to understand the unified interface.
• Concerned about costs? Go straight to Cost-Aware LLM Router.
• Building production services? Read LLM Fallback Chains and Semantic Response Caching.
• Need structured data from LLMs? See Structured Output with Pydantic.