POST/api/v1/pipelines/run

Run pipeline

Primary native entry: prompt, provider, model, optional modules (cache, RAG, compression, PII), context layers, and streaming flag. Same contract as the console pipeline.

Required fields include a text prompt and a catalogue-backed provider/model pair. Optional booleans and thresholds mirror the Pipeline IA settings in the console.

Python SDK: TokenSaver.ask() and run_pipeline() POST this endpoint. See sdk/python/pipelines for parameter tables.

bash

curl -sS "$API/api/v1/pipelines/run" \

  -H "Authorization: Bearer $TS_KEY" \

  -H "Content-Type: application/json" \

  -d '{

    "prompt": "Hello",

    "provider": "openai",

    "model": "gpt-4o",

    "use_cache": true

}'