Skip to content

START HERE - Traigent Examples Navigator

Pick an example in under a minute.

Quick picks

  • 2 min: python examples/core/simple-prompt/run.py
  • 5 min: TRAIGENT_MOCK_LLM=true python examples/core/rag-optimization/run.py
  • 8 min: TRAIGENT_MOCK_LLM=true python examples/core/few-shot-classification/run.py
  • 10 min: TRAIGENT_MOCK_LLM=true python examples/core/multi-objective-tradeoff/run_anthropic.py
  • 30+ min: browse examples/advanced/ playbooks

Choose by goal

  • Lower cost/latency: multi-objective tradeoff, token-budget summarization, execution-modes (advanced).
  • Better accuracy: few-shot classification, prompt-style optimization, AI engineering tasks (few-shot selection).
  • Safety: safety-guardrails, AI engineering tasks (safety).
  • Structured JSON: structured-output-json, AI engineering tasks (structured output).
  • Tools: tool-use-calculator.
  • Long docs: chunking-long-context, AI engineering tasks (context engineering).
  • Prompt tests: prompt-ab-test.
  • CI/CD: examples/integrations/ci-cd/.

Directory map (current)

examples/
|- core/
|  |- simple-prompt/
|  |- rag-optimization/
|  |- few-shot-classification/
|  |- multi-objective-tradeoff/
|  |- token-budget-summarization/
|  |- structured-output-json/
|  |- tool-use-calculator/
|  |- prompt-style-optimization/
|  |- chunking-long-context/
|  |- safety-guardrails/
|  `- prompt-ab-test/
|
|- advanced/
|  |- execution-modes/
|  |- results-analysis/
|  |- ai-engineering-tasks/
|  |- ragas/
|  `- metric-registry/
|
|- integrations/
|  |- ci-cd/
|  `- bedrock/
|
|- datasets/
|- gallery/
|- templates/
|- utils/
`- tvl/
  1. core/simple-prompt/ - sanity check and mock mode setup
  2. core/rag-optimization/ - RAG toggle and telemetry
  3. core/few-shot-classification/ - few-shot concepts
  4. core/multi-objective-tradeoff/ - balance accuracy, latency, cost
  5. advanced/ai-engineering-tasks/ - specialized playbooks

Run instructions

cd /path/to/Traigent
pip install "traigent[recommended]"
python examples/core/simple-prompt/run.py

With real APIs (optional): 

export ANTHROPIC_API_KEY="your-key" # pragma: allowlist secret
export OPENAI_API_KEY="your-key" # pragma: allowlist secret
python examples/core/multi-objective-tradeoff/run_anthropic.py

Pro tips

  • Stay in repo root so relative paths resolve.
  • Use mock mode first to avoid spend.
  • Every example notes its dataset under examples/datasets/<example>/.
  • Copy an example folder into your project to bootstrap a workflow.

Need more? See Examples Guide or Troubleshooting.