# Sample Raw Data Index / 原始数据索引 (mock)

> **MOCK SAMPLE — sanitized public sample, not a paid customer report.**
> Audit: `ord_sample_snapshot_20260615` · Sampled: `2026-06-15T10:00:00+08:00` · Product: `snapshot`
> Run scope: 5 user queries × 7 endpoints = 35 datapoints
> Public base URL: `example.test/2026-06-15/ord_sample_snapshot_20260615/` (placeholder — does not resolve)
> Subject brand: `Example Cross-Border Brand` (URL `example.com`, fictional per IANA reserved domains)
> See <https://clarivy.ai/audit/sample/> for the full disclaimer.

This file is the canonical index of every raw JSON that powers the AI-readable Markdown brief and the human PDF. Every finding
in the brief traces back to one of the rows below. In a paid run this index points to a dated folder in the public
`agentgeek-geo/audit-logs` repo; in this public sample the URLs are placeholders under `example.test` and **do not
resolve** — they exist only to show the deliverable structure.

## Per-endpoint raw JSON count

| endpoint | datapoints | errors | path |
|---|---|---|---|
| doubao | 5 | 0 | `example.test/2026-06-15/ord_sample_snapshot_20260615/doubao/` |
| kimi | 5 | 0 | `example.test/2026-06-15/ord_sample_snapshot_20260615/kimi/` |
| deepseek | 5 | 0 | `example.test/2026-06-15/ord_sample_snapshot_20260615/deepseek/` |
| qwen | 5 | 0 | `example.test/2026-06-15/ord_sample_snapshot_20260615/qwen/` |
| chatgpt | 5 | 0 | `example.test/2026-06-15/ord_sample_snapshot_20260615/chatgpt/` |
| claude | 5 | 0 | `example.test/2026-06-15/ord_sample_snapshot_20260615/claude/` |
| gemini_aio | 5 | 0 | `example.test/2026-06-15/ord_sample_snapshot_20260615/gemini_aio/` |

## Per-prompt × per-endpoint mention grid

> 1 = brand mentioned · 0 = brand not mentioned · — = datapoint missing/error

| prompt_id | category | text | `doubao` | `kimi` | `deepseek` | `qwen` | `chatgpt` | `claude` | `gemini_aio` |
|---|---|---|---|---|---|---|---|---|---|
| U01 | brand_agnostic_category | best GEO audit service for cross-border brands in 2026 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| U02 | brand_specific | is example.com legit | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| U03 | brand_specific_competitive | example.com vs Profound | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| U04 | category_comparison | best AI search visibility audit for DTC brand | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| U05 | improvement_intent | how to improve example.com citations in AI answers | 0 | 0 | 0 | 0 | 0 | 0 | 0 |

## Per-prompt raw JSON files

> In a paid run, each row below resolves to a public JSON file you can open in your browser. In this public sample the
> URLs are placeholders under `example.test` and **do not resolve**.

### U01 (best GEO audit service for cross-border brands in 2026)

- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/doubao/U01.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/kimi/U01.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/deepseek/U01.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/qwen/U01.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/chatgpt/U01.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/claude/U01.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/gemini_aio/U01.json` (brand: 0)

### U02 (is example.com legit)

- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/doubao/U02.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/kimi/U02.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/deepseek/U02.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/qwen/U02.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/chatgpt/U02.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/claude/U02.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/gemini_aio/U02.json` (brand: 0)

### U03 (example.com vs Profound)

- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/doubao/U03.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/kimi/U03.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/deepseek/U03.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/qwen/U03.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/chatgpt/U03.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/claude/U03.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/gemini_aio/U03.json` (brand: 0)

### U04 (best AI search visibility audit for DTC brand)

- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/doubao/U04.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/kimi/U04.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/deepseek/U04.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/qwen/U04.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/chatgpt/U04.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/claude/U04.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/gemini_aio/U04.json` (brand: 0)

### U05 (how to improve example.com citations in AI answers)

- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/doubao/U05.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/kimi/U05.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/deepseek/U05.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/qwen/U05.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/chatgpt/U05.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/claude/U05.json` (brand: 0)
- `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/gemini_aio/U05.json` (brand: 0)

## How to use this index

1. Pick a `finding_id` (e.g. `F-001-total-coverage`) from the AI-readable Markdown brief.
2. Find the corresponding `evidence_id` (e.g. `E-DOUBAO-U01`) in Section 4 of the brief.
3. Open the raw JSON at `https://clarivy.ai/audit-logs/2026-06-15/<order_id>/raw/<endpoint>/<prompt_id>.json` (paid run) — in this
   public sample, the URL is the placeholder `example.test/2026-06-15/ord_sample_snapshot_20260615/raw/<endpoint>/<prompt_id>.json`
   and **does not resolve**.

## Provenance

- Who ran: agentgeek-geo/audit-runner orchestrator (mock mode, `RUN_MODE=mock`).
- Where: 7 endpoints (Doubao / Kimi / DeepSeek / Qwen / ChatGPT / Claude / Gemini) — see <https://clarivy.ai/audit/methodology.html#production-endpoints>.
- When: 2026-06-15T10:00:00+08:00.
- Total datapoints: 35.
- Errors: 0 (all 35 contract datapoints returned non-empty, non-error responses — `confidence: high` per
  <https://clarivy.ai/audit/methodology.html#confidence-accuracy-limitations>).

---

**Sample disclaimer.** This file is a sanitized public sample, not a paid customer report. The subject brand is
fictional, the 5 user queries and 35 datapoints are mock, and every URL is a placeholder. For the full disclaimer
see <https://clarivy.ai/audit/sample/>. License: CC-BY-4.0 (attribute clarivy.ai if reused).