← All skills
Skill · Prompting → chains → RAG → agentsLive

LLM Apps

The patterns behind every AI app — prompt templates, chaining, structured output, retrieval, and tool-using agents — wired to a real, live model. No keys, no setup.

  1. Task 1Your first callEasySend text to a real model and read its reply.~10 min · 1 blanks→
  2. Task 2Prompt templatesEasyBuild prompts from a reusable f-string template.~11 min · 1 blanks→
  3. Task 3System promptsEasySet the model's rules and persona with a system message.~11 min · 1 blanks→
  4. Task 4Few-shot promptingMediumTeach the model a pattern by showing examples first.~12 min · 1 blanks→
  5. Task 5Structured JSON outputMediumGet machine-readable JSON instead of prose.~13 min · 1 blanks→
  6. Task 6Chaining callsMediumFeed one model's output into the next call.~13 min · 1 blanks→
  7. Task 7EmbeddingsMediumTurn text into a vector of numbers.~12 min · 1 blanks→
  8. Task 8Semantic searchHardRank documents by meaning, not keywords.~14 min · 1 blanks→
  9. Task 9RAG: answer from docsHardGround the model in facts so it can't make things up.~15 min · 1 blanks→
  10. Task 10A tool-using agentHardLet the model choose a tool; your code runs it.~16 min · 1 blanks→