Be Seen.
Be Cited.
Be Chosen.
搜索时代问「我排在哪里」。AI 时代问「我被如何记住」。
SeenGeo 是中文品牌在 ChatGPT、DeepSeek、豆包、通义、Kimi 中的镜像。
loading mirrors…
01 / Brands
搜索时代问「我排在哪里」。AI 时代问「我被如何记住」。
SeenGeo 是中文品牌在 ChatGPT、DeepSeek、豆包、通义、Kimi 中的镜像。
loading mirrors…
被搜索的时代已经结束。 现在的问题是: 你被怎样记住。
这是中文品牌在 5 个 LLM 中被记住的形状,每天更新一次,公开只读。
5 endpoints·GET only·JSON·rate limit 60/min
M1 阶段所有公开读端点免认证。客户端按 IP 限流 60 req/min,超限返回 429 与 Retry-After header。CORS 允许所有来源(Access-Control-Allow-Origin: *)。
All public read endpoints are unauthenticated in M1. Per-IP rate limit is 60 req/min — over limit returns 429 with a Retry-After header.
/api/v1/brands返回品牌池中全部(或按 category 过滤)的品牌,带聚合统计:扫描次数、最近扫描时间、综合评分。
categoryquery品牌大类,如 new_consumer / friendslimitquery返回数量上限(1-200)curl https://seengeo.com/api/v1/brands?category=new_consumer&limit=50
{
"count": 30,
"brands": [
{
"id": "heytea",
"cn": "喜茶",
"en": "HEYTEA",
"category": "new_consumer",
"sub_category": "tea_drinks",
"bucket": "priority",
"hq": "shenzhen",
"scan_count": 124,
"last_scanned_at": "2026-04-27T03:14:00.000Z",
"score_total": 0.71
}
]
}/api/v1/brands/{brand}返回单个品牌的完整元信息、最新评分摘要(按 5 维度聚合)、扫描计数。
brandpath · required品牌 slug,如 heyteacurl https://seengeo.com/api/v1/brands/heytea
{
"meta": {
"id": "heytea",
"cn": "喜茶",
"en": "HEYTEA",
"category": "new_consumer",
"sub_category": "tea_drinks",
"competitors": ["nayuki", "luckin"],
"signature_products": ["多肉葡萄", "芝士奶盖"]
},
"scan_count": 124,
"engine_count": 5,
"last_scanned_at": "2026-04-27T03:14:00.000Z",
"latest_scores": {
"total": 0.71,
"by_dimension": {
"facts_accuracy": { "avg": 0.82, "count": 24, "confidence": 0.9 },
"unprompted_top3_hit": { "avg": 0.66, "count": 12, "confidence": 0.8 }
}
}
}/api/v1/brands/{brand}/scores返回最近 N 天的全部维度评分记录,每条 (scan, dimension) 一行。包含评分理由文本、置信度、判官模型。
brandpath · required品牌 slugdaysquery回溯天数(默认 30,最大 30)curl https://seengeo.com/api/v1/brands/heytea/scores?days=30
{
"brand_id": "heytea",
"days": 30,
"count": 612,
"scores": [
{
"score_id": "01HX...",
"scan_id": "01HX...",
"engine_id": "gpt-5.5",
"question_id": "q3_unprompted_recommendation",
"dimension": "unprompted_top3_hit",
"score": 0.67,
"confidence": 0.82,
"reasoning": "排在第 2 位被推荐,理由完整。",
"judge_model": "claude-opus-4-7",
"scored_at": "2026-04-27T03:14:00.000Z"
}
]
}/api/v1/brands/{brand}/scans返回品牌的扫描记录,含 LLM 原始回答前 280 字符摘要、token 用量、延迟。可按 question_id 过滤。
brandpath · required品牌 slugquestionquery问题 ID,如 q3_unprompted_recommendationdaysquery回溯天数(默认 7,最大 30)limitquery上限(默认 50,最大 200)curl 'https://seengeo.com/api/v1/brands/heytea/scans?question=q3_unprompted_recommendation&days=7&limit=50'
{
"brand_id": "heytea",
"days": 7,
"question": "q3_unprompted_recommendation",
"count": 35,
"scans": [
{
"scan_id": "01HX...",
"engine_id": "gpt-5.5",
"question_id": "q3_unprompted_recommendation",
"status": "ok",
"response_excerpt": "如果想买茶饮,我会推荐:1. 喜茶 — ...",
"tokens_in": 142,
"tokens_out": 380,
"latency_ms": 2418,
"started_at": "2026-04-27T03:14:00.000Z",
"finished_at": "2026-04-27T03:14:02.418Z"
}
]
}/api/v1/anomalies对比窗口前半段 vs 后半段的评分均值,返回 |Δ| ≥ 0.2 的维度异动。Severity 分 low(≥0.2)/ medium(≥0.4)/ high(≥0.6)。
hoursquery窗口小时数(默认 24,最大 720)severityquery过滤 high / medium / lowcurl 'https://seengeo.com/api/v1/anomalies?hours=24&severity=high'
{
"window_hours": 24,
"severity": "high",
"count": 2,
"anomalies": [
{
"brand_id": "heytea",
"dimension": "unprompted_top3_hit",
"dimension_cn": "无提示推荐命中",
"prev_score": 0.78,
"current_score": 0.12,
"delta": -0.66,
"hours": 24,
"severity": "high"
}
]
}完整规范以 JSON 提供,可直接喂给 Postman / Stoplight / 任何代码生成器: