LLM

Foundation

General-purpose clinical language model inference on AWS Bedrock.

LLMBedrockCompletion

About

Direct access to the platform's language model for free-form generation: summarization, question answering, classification, and drafting. Runs on AWS Bedrock via the converse() API with the model selected by platform infrastructure.

This is the escape hatch when no specialized service fits. For coding, auditing, prior auth, or redaction, the dedicated services add multi-pass validation, evidence grounding, and governance that a raw completion cannot provide.

How it works

  1. 1Prompt + optional system prompt submitted as JSON
  2. 2Bedrock converse() call with configured temperature / max_tokens (extended thinking optional)
  3. 3Response text returned with normalized token_usage for cost tracking

Intended use

  • •Summarizing extracted document text or structured results
  • •Custom classification and Q&A steps inside generated apps
  • •Glue logic between specialized services (e.g. turning audit findings into a patient-friendly letter)

Key outputs

  • ▸response — the generated completion text
  • ▸token_usage — prompt/completion/total/cached token counts from Bedrock

Endpoints

Try each endpoint with your signed-in session — usage counts toward your monthly budget.

Use synthetic data only. Do not submit real patient records or PHI when testing endpoints.

Limitations & caveats

  • –Do not send a model field — the healthcare API selects the model internally; client-specified models are ignored
  • –No evidence grounding or governance: outputs are unvalidated free text, unsuitable as a substitute for the coding/audit pipelines
  • –Single-turn completion endpoint — conversation state must be managed by the caller