DataForSEO Brain

Content Analysis API (Capability)

4 min read

Content Analysis API (Capability)

The Content Analysis API searches a global citation database of web mentions for a keyword or category, then returns the citations, their sentiment and rating distributions, and how mention volume and phrasing trend over time. Sits under DataForSEO Brain → Concepts.

Overview

This module is brand-mention and sentiment intelligence over a pre-built citation corpus, aggregating content from news sites, ecommerce, blogs, forums, and organization pages. You query it like a search engine for mentions: give it a keyword (or a category code) and it returns every place that term is cited, each scored for sentiment and connotation, plus aggregate summaries, rating distributions, and time-series trends. All endpoints are Live/synchronous POST calls under /v3/content_analysis/{function}/live; there is no task queue. It is the listening layer for reputation tracking and content-gap research.

What it covers

  • /v3/content_analysis/search/live - detailed citation data for a keyword: snippet, sentiment, domain metrics, page categorization.
  • /v3/content_analysis/summary/live - an aggregate overview (sentiment, top domains, categorical breakdowns) of the citation set.
  • /v3/content_analysis/sentiment_analysis/live - emotional-reaction and polarity distribution for a keyword.
  • /v3/content_analysis/rating_distribution/live - rating-distribution metrics across domains, categories, and regions.
  • /v3/content_analysis/phrase_trends/live and /v3/content_analysis/category_trends/live - citation statistics over time for a keyword or category (history from 2022-10-31).
  • Free helpers: /v3/content_analysis/categories/, GET /v3/content_analysis/available_filters (the filterable-fields reference; the /filters/ form 404s for this module), and /v3/content_analysis/id_list/.

Key parameters / inputs

fieldapplies tonotes
keywordsearch, summary, sentiment_analysis, rating_distribution, phrase_trendstarget term, UTF-8, lowercased; quotes for exact phrase
category_codecategory_trends (ids come from the categories helper)target category id
keyword_fieldssearch, summaryrestrict to title / main_title / previous_title / snippet
page_typesearch, summaryecommerce / news / blogs / message-boards / organization
search_modesearch, summaryas_is (all citations) or one_per_domain
positive_connotation_threshold / sentiments_connotation_thresholdsummary, sentiment_analysis (NOT search)sentiment thresholds 0-1; default 0.4
date_from / date_to / date_groupphrase_trends, category_trends onlytrends window (min 2022-10-31), grouped by day/week/month
filters / internal_list_limit / limit / offset_tokensearch and other list-returning endpointsmax 8 filters, nested-array cap, limit up to 1000, token pagination beyond 10000

Response / what you get back

Search returns total_count, items_count, offset_token, and per-citation type, url, domain, main_domain, url_rank, domain_rank, content_info.title, content_info.snippet, content_info.sentiment_connotations, content_info.connotation_types, content_info.rating, page_category, and fetch_time. The sentiment object holds anger, happiness, love, sadness, share, and fun; the connotation object holds positive, negative, and neutral. Aggregate endpoints add top_domains, text_categories, page_categories, page_types, countries, and languages; rating distribution adds bucketed rating_distribution; trends add a date series.

Cost & method notes

  • All analysis endpoints are Live; per-request billing. The cost log shows /v3/content_analysis/search/live at 0.02003.
  • The categories, filters, and id_list helpers are free.
  • Because the corpus is pre-indexed, costs are flat per request rather than per result; see cap-queue-priority-cost-model.

When to use / how it fits

Use Content Analysis to gauge how a brand, product, or topic is talked about and where the content gaps are. It feeds the sentiment step of play-content-strategy-brief and complements review intelligence in cap-business-data-api and plat-review-platforms. Its category taxonomy mirrors cap-labs-market-analysis, so the two pair well for demand-plus-sentiment mapping. Route jobs via dec-which-api-for-which-job.

Gotchas / limits

  • Trend history starts 2022-10-31; there is nothing earlier.
  • All endpoints are synchronous only, with no POST-then-GET queue; id_list is rate-limited to 10 calls/min and returns task IDs up to 6 months back, but metadata is only available for tasks from the past month, and include_metadata: true caps the lookback to 1 month.
  • 2000 API calls/min, max 8 filters, max 3 sort rules; use offset_token past 10000 results.
  • Sentiment is model-derived from text, not verified human labels; treat thresholds as tunable heuristics.
  • search_mode: one_per_domain collapses many citations from the same site into a single row, which changes both total_count and the apparent sentiment mix versus as_is.
  • The keyword_fields and page_type filters scope where the term is matched (title vs snippet, news vs ecommerce vs forum), so the same keyword can yield very different citation sets depending on how you constrain it.

Sources

Graph View

Backlinks