---
title: "Grade Session"
method: POST
path: "/sessions/{session_id}/grade"
---

# Grade Session

`POST /sessions/{session_id}/grade`

Grade a single session against ground truth using LLM judge.

The trace is fetched from S3 (environment-agnostic), so sessions from
any deployment (e.g. staging) can be graded. Platform-admin auth is
the access control; no per-session org check is needed.

## Path parameters

- `session_id` string, required

## Request body

- GradeSessionRequest — Request to grade a single session against ground truth.
  - `ground_truth` string, required — Ground truth text to grade against
  - `dimensions` GradingDimension[], nullable — Custom grading dimensions. Uses defaults if omitted.
    - `name` string, required — Name/identifier for this grading dimension
    - `criteria` string, required — Evaluation criteria including the 1–5 scale description
    - `priority` integer — Priority for display ordering (higher = first)

## Response `200`

Successful Response

- GradeSessionResponse — Response from grading a session.
  - `overall_score` number, required
  - `dimension_scores` object, required
  - `justifications` object, required

## Other responses

- `422` — Validation Error

---

[API](https://skmtc.net/traversal/apis/fastapi.md) · [All operations](https://skmtc.net/traversal/apis/fastapi/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/traversal/fastapi/revisions/2134ebffd1ef/schema)
