> ## Documentation Index
> Fetch the complete documentation index at: https://docs.medera.info/llms.txt
> Use this file to discover all available pages before exploring further.

# Facts & Reasoning

> Structured facts extraction

Medera's facts pipeline extracts structured `Fact` objects from the encounter — every clinically relevant claim, who said it, and supporting evidence. Documents generate against the fact set rather than raw transcripts so every claim ties back to a source.

The real-time fact stream rides the same WebSocket as the transcript on the clinical session and intake pipelines.

```text theme={"system"}

POST /api/therapy/analyze-session   # extracts facts + downstream synthesis

```

A `Fact` includes:

```json theme={"system"}
{
  "fact_id": "fact_abc",
  "group": "presenting_problem",
  "claim": "Persistent low mood for 6 weeks",
  "evidence_spans": ["span_004"],
  "source_speaker": "patient",
  "certainty": "asserted",
  "timestamp_ms": 84300
}

```