---
title: "Reasoning Pregrade"
method: POST
path: "/api/asclepius/reasoning/pregrade"
tags: ["asclepius"]
---

# Reasoning Pregrade

`POST /api/asclepius/reasoning/pregrade`

## Request

- Base URL: `https://app.archangelhealth.ai` (the document declares no server; this is the origin it was published from)
- URL: `https://app.archangelhealth.ai/api/asclepius/reasoning/pregrade`
- Auth: none declared

Split + pre-grade an answer's reasoning steps (Speed Optimization §2):
each step arrives with a SUGGESTED ``good``/``bad`` label (+ a one-line
critique on bad steps) so the doctor verifies instead of authoring. The
labels are suggestions only — every step still requires an explicit human
confirm/correct before submit. Degrades to the heuristic splitter with
``suggested_label = null`` when no LLM is configured.

## Headers

- `authorization` string, nullable

## Request body

- ReasoningSplitRequest: Split a chosen/ideal answer into ordered reasoning steps for tap-to-grade (Eval Flow Upgrade §4). ``prompt``/``specialty`` give the splitter context.
  - `text` string, required
  - `prompt` string
  - `specialty` string

## Response `200`

Successful Response

- unknown

## Other responses

- `422`: Validation Error

---

[API](https://skmtc.dev/archangelhealth/apis/archangel-health-surgical-patient-platform.md) - [All operations](https://skmtc.dev/archangelhealth/apis/archangel-health-surgical-patient-platform/llms.txt) - [OpenAPI document](https://skmtc.dev/archangelhealth/apis/archangel-health-surgical-patient-platform/revisions/3514f4e119da?raw)