Obtener respuestas

curl -X POST https://api.tryprofound.com/v2/prompts/answers \
  -H "X-API-Key: <your-api-key>" \
  -H "Content-Type: application/json" \
  -d '{
    "category_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
    "start_date": "2026-06-18",
    "end_date": "2026-06-19",
    "limit": 10
  }'

{
  "info": {
    "total_results": 27690,
    "count": 10,
    "next_cursor": "Z0FBQUFBQnFQcjNsbDZZaWlzZjR2cXJt...",
    "models": ["ChatGPT", "Google Gemini", "Perplexity", "Claude", "..."],
    "include": [
      "run_id", "date", "model", "topic", "topic_id", "persona", "region",
      "tags", "prompt", "prompt_id", "response", "mentions", "citations",
      "search_queries", "analysis_types", "sentiment_claims"
    ],
    "start_date": "2026-06-18",
    "end_date": "2026-06-19",
    "filter": null
  },
  "data": [
    {
      "run_id": "<uuid>",
      "date": "2026-06-19",
      "model": { "id": "<uuid>", "name": "ChatGPT" },
      "topic": "<string>",
      "topic_id": "<uuid>",
      "region": "United States",
      "persona": null,
      "tags": [],
      "prompt": "<prompt text>",
      "prompt_id": "<uuid>",
      "response": "<full model response text>",
      "mentions": ["Profound", "<string>"],
      "citations": ["https://example.com/article"],
      "search_queries": ["<search query>"],
      "analysis_types": ["visibility"],
      "sentiment_claims": [
        { "asset": "Profound", "theme": "<string>", "claim": "<string>", "sentiment": "positive" }
      ]
    }
    // ...9 more rows (count: 10, total_results: 27690); paginate with next_cursor
  ]
}

POST

prompts

answers

curl -X POST https://api.tryprofound.com/v2/prompts/answers \
  -H "X-API-Key: <your-api-key>" \
  -H "Content-Type: application/json" \
  -d '{
    "category_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
    "start_date": "2026-06-18",
    "end_date": "2026-06-19",
    "limit": 10
  }'

{
  "info": {
    "total_results": 27690,
    "count": 10,
    "next_cursor": "Z0FBQUFBQnFQcjNsbDZZaWlzZjR2cXJt...",
    "models": ["ChatGPT", "Google Gemini", "Perplexity", "Claude", "..."],
    "include": [
      "run_id", "date", "model", "topic", "topic_id", "persona", "region",
      "tags", "prompt", "prompt_id", "response", "mentions", "citations",
      "search_queries", "analysis_types", "sentiment_claims"
    ],
    "start_date": "2026-06-18",
    "end_date": "2026-06-19",
    "filter": null
  },
  "data": [
    {
      "run_id": "<uuid>",
      "date": "2026-06-19",
      "model": { "id": "<uuid>", "name": "ChatGPT" },
      "topic": "<string>",
      "topic_id": "<uuid>",
      "region": "United States",
      "persona": null,
      "tags": [],
      "prompt": "<prompt text>",
      "prompt_id": "<uuid>",
      "response": "<full model response text>",
      "mentions": ["Profound", "<string>"],
      "citations": ["https://example.com/article"],
      "search_queries": ["<search query>"],
      "analysis_types": ["visibility"],
      "sentiment_claims": [
        { "asset": "Profound", "theme": "<string>", "claim": "<string>", "sentiment": "positive" }
      ]
    }
    // ...9 more rows (count: 10, total_results: 27690); paginate with next_cursor
  ]
}

Las filas sin procesar por ejecución que hay detrás de cada informe: una fila por respuesta del modelo, con sus menciones, citas y consultas de búsqueda. Sin agregación.

Sin group_by / sort: las filas se ordenan por fecha, de la más reciente a la más antigua.
include elige qué campos devolver (response, mentions, citations, search_queries, sentiment_claims, …); omítelo para obtener todos los campos excepto citation_details, que debe solicitarse explícitamente.
mentions es una lista plana y sin duplicados, en el orden en que aparecen las entidades en la respuesta.
sentiment_claims enumera las afirmaciones extraídas de la respuesta, cada una con asset, theme, claim y sentiment (positive o negative).
Filtros: campos a nivel de prompt más analysis_type (a nivel de prompt: visibility·sentiment·factcheck), y hojas domain/page de nivel superior: is un valor o in una lista (coincidencia exacta de la URL citada).

¿Es la primera vez que usas los informes v2? Consulta Filtrado y conceptos para conocer la estructura común de las solicitudes, el árbol de filtros, la agrupación y la paginación.

Variante de streaming (SSE) (mismo cuerpo, /stream)

POST /v2/prompts/answers/stream acepta el mismo cuerpo de solicitud y devuelve Server-Sent Events: un evento summary (el bloque info) y luego un evento result por fila. limit/cursor se ignoran; de forma predeterminada devuelve todo. Pasa max_results para establecer un límite.

Response (text/event-stream)

event: summary
data: { ...the info block... }

event: result    ← one per row, same shape as data[] above
data: {"date": "2026-06-19", "model": {"name": "ChatGPT"}, "prompt": "<prompt text>", "mentions": ["Profound", "<string>"], "citations": ["https://example.com/article"]}

curl -X POST https://api.tryprofound.com/v2/prompts/answers \
  -H "X-API-Key: <your-api-key>" \
  -H "Content-Type: application/json" \
  -d '{
    "category_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
    "start_date": "2026-06-18",
    "end_date": "2026-06-19",
    "limit": 10
  }'

{
  "info": {
    "total_results": 27690,
    "count": 10,
    "next_cursor": "Z0FBQUFBQnFQcjNsbDZZaWlzZjR2cXJt...",
    "models": ["ChatGPT", "Google Gemini", "Perplexity", "Claude", "..."],
    "include": [
      "run_id", "date", "model", "topic", "topic_id", "persona", "region",
      "tags", "prompt", "prompt_id", "response", "mentions", "citations",
      "search_queries", "analysis_types", "sentiment_claims"
    ],
    "start_date": "2026-06-18",
    "end_date": "2026-06-19",
    "filter": null
  },
  "data": [
    {
      "run_id": "<uuid>",
      "date": "2026-06-19",
      "model": { "id": "<uuid>", "name": "ChatGPT" },
      "topic": "<string>",
      "topic_id": "<uuid>",
      "region": "United States",
      "persona": null,
      "tags": [],
      "prompt": "<prompt text>",
      "prompt_id": "<uuid>",
      "response": "<full model response text>",
      "mentions": ["Profound", "<string>"],
      "citations": ["https://example.com/article"],
      "search_queries": ["<search query>"],
      "analysis_types": ["visibility"],
      "sentiment_claims": [
        { "asset": "Profound", "theme": "<string>", "claim": "<string>", "sentiment": "positive" }
      ]
    }
    // ...9 more rows (count: 10, total_results: 27690); paginate with next_cursor
  ]
}

Autorizaciones

X-API-Key

string

header

requerido

Cuerpo

application/json

category_id

string<uuid>

requerido

start_date

string

requerido

YYYY-MM-DD, ET, inclusive

end_date

string

requerido

YYYY-MM-DD, ET, inclusive

include

enum<string>[] | null

Which row fields to return: run_id, date, model, topic, topic_id, region, persona, tags, prompt, prompt_id, response, mentions, citations, citation_details, search_queries, analysis_types, sentiment_claims. Omit for all fields except citation_details, which must be requested explicitly because it is expensive.

Opciones disponibles:

run_id,

date,

model,

topic,

topic_id,

persona,

region,

tags,

prompt,

prompt_id,

response,

mentions,

citations,

citation_details,

search_queries,

analysis_types,

sentiment_claims

filter

FilterNode · object | null

and/or/not tree over model, topic, region, persona, prompt, tag, analysis_type (visibility/sentiment/factcheck); plus top-level and leaves domain or page (is one value, or in a list). Substring-search the prompt with {"field": "prompt", "op": "contains", "value": "…"}.

Show child attributes

limit

integer | null

Page size; default 10, max 200.

Rango requerido: 0 < x <= 200

max_results

integer | null

Stream endpoint only: cap the number of streamed rows (default: all).

Rango requerido: x > 0

cursor

string | null

Respuesta

Successful Response

info

AnswersV2Info · object

requerido

Show child attributes

data

AnswerRow · object[]

requerido

Show child attributes

Descripción general

SDKs

Integraciones

Agentes

Organización

Categorías

Prompts

Bases de conocimiento

OpenAI Ads

Respuestas

Volúmenes de prompts

Informes

Optimización de contenido

Obtener respuestas

Autorizaciones

Cuerpo

Respuesta