Definition
Full definition of llm
LLMs take text in, produce text out. The "large" part means they have billions or trillions of parameters trained on internet-scale text. Useful for classification, summarization, drafting, code generation, and reasoning. Tiny Command has first-class actions for OpenAI, Anthropic (Claude), and other LLM providers. Most AI-driven automation flows you'll build use an LLM somewhere in the middle.
In practice
LLM examples
LLM call
Prompt: 'Summarize this email in 1 sentence: [...]' → Output: 'The customer wants to upgrade to Pro.'
Used by
Apps that exemplify llm
See llm in action across real integrations.
FAQ
Common questions about llm
Claude vs GPT vs Gemini?
All three are competent. Claude shines for long-context reasoning; GPT for ecosystem maturity; Gemini for multimodal + Google integration. Test on your real workload before standardizing.
How do I cost-control LLM calls?
Pick the smallest model that works (gpt-4o-mini, Claude Haiku). Cap max_tokens. Cache responses where possible. Track per-call cost from the usage field.
