Apify Google AI Overview API

Google AI Overview API: Answers and Citations as JSON

Fetch Google's AI Overview for any search query as clean, structured JSON. One call takes a query (or a batch of queries) and returns the AI-generated answer as ordered text blocks plus every source it cites, with links and source domains.

When to use this skill

• The user wants a Google AI Overview API, or Google AI Overview data programmatically.
• They want to extract or export AI Overview answers to JSON or CSV.
• They want the cited sources behind an AI Overview for a keyword or topic.
• They want to batch AI Overview lookups across a keyword list into one dataset.

Not for: recurring brand-citation watchlists (use the google-ai-overview-monitoring skill), organic result rank tracking (this returns the AI Overview only), or other engines' AI answers (see references/actor-index.md for Bing, Brave, and Naver equivalents).

What you get (one row per query)

query, result_type, ai_overview_present, text_blocks (ordered blocks typed paragraph, heading, list, or table, each with a snippet; list blocks carry list items, table blocks carry raw table rows plus a formatted object form), snippet_links and snippet_highlighted_words when present, reference_indexes tying a block to its sources when present, references (each with link, source, index, plus title, snippet, thumbnail, and source_icon when present), used_dedicated_api, retrievals_used, gl, hl, fetched_at.

Prerequisites

• Apify account (sign up at https://apify.com?fpr=9n7kx3&fp_sid=skillrepo).
• Authentication via apify login, or an APIFY_TOKEN environment variable (Apify Console, Settings, Integrations).

The Actor

• Store page: https://apify.com/johnvc/Google-AI-Overview-API?fpr=9n7kx3&fp_sid=skillrepo
• Actor ID: johnvc/Google-AI-Overview-API
• Pricing: pay per AI Overview retrieval plus a small per-run setup fee (see references/gotchas.md).

Run it with the Apify CLI

Fetch one AI Overview:

apify actors call "johnvc/Google-AI-Overview-API" -i '{"query":"what is retrieval augmented generation"}' \
  --json \
  --user-agent apify-awesome-skills/apify-google-ai-overview-api \
  2>/dev/null

Batch a keyword list with country targeting:

apify actors call "johnvc/Google-AI-Overview-API" -i '{"queries":["best crm for startups","how does vector search work","what is a transformer model"],"gl":"us","hl":"en"}' \
  --json \
  --user-agent apify-awesome-skills/apify-google-ai-overview-api \
  2>/dev/null

Every call carries the three flags this repo expects: --json, --user-agent apify-awesome-skills/apify-google-ai-overview-api, and 2>/dev/null.

Run it from Claude or another AI agent (MCP)

The Actor is MCP-ready. Add the hosted server URL:

https://mcp.apify.com/?tools=actors,docs,johnvc/Google-AI-Overview-API

Then ask, for example: "Get the Google AI Overview for 'best crm for startups' and list the sources it cites." MCP setup docs: https://docs.apify.com/platform/integrations/mcp

Workflow

1. Collect the queries. Provide a single query, a queries list, or both; they merge and dedupe, and the Actor returns one row per query. Question-style informational searches trigger AI Overviews most often.
2. Set targeting. Keep hl at en (AI Overviews are currently English-only), pick gl for the country, and add location only when a named place matters.
3. Estimate cost, then confirm with the user if the batch is large. See references/gotchas.md.
4. Run the Actor and read the dataset. To get plain answer text, concatenate the snippet of each block in text_blocks in order. To attribute claims, map each block's reference_indexes to the references array by index.
5. Deliver rows as JSON or CSV, or hand back the dataset link. fetched_at timestamps each row for later comparisons.

Inputs

• query (string): a single search query
• queries (array of strings): batch of queries, merged with query and deduped
• At least one of query or queries is required
• gl (string, default us): two-letter country code
• hl (string, default en): interface language; keep en
• location (string): optional named location, for example "Austin, Texas, United States"

Cost

Billing is pay per event: a small setup fee per run plus a fee per AI Overview retrieval. An inline answer costs one retrieval; when Google defers generation, the Actor resolves it automatically at the cost of a second retrieval. A 10-query batch of inline answers runs about $0.16. Live prices and thresholds are in references/gotchas.md.

Honest limits

• Not every query gets an AI Overview. Rows without one come back with ai_overview_present: false; you still pay for the lookup.
• English-only for now, and available in a limited set of countries; hl other than en will mostly return nothing.
• Some references entries carry only link, source, and index (video sources often skip title and snippet).
• Answers change over time and vary by gl and location; this fetches the live overview, not history. Build history with your own repeated runs.

Troubleshooting

• ai_overview_present: false: rephrase as a question-style informational query, keep hl at en, and try gl at us.
• used_dedicated_api: true and retrievals_used: 2: normal; Google deferred generation and the Actor resolved it, which costs one extra retrieval.
• Missing title or snippet on a reference: normal for some source types; link and source are always there for matching.

See references/gotchas.md for cost guardrails and error recovery, and references/actor-index.md for the Actor routing table.

Related answer-engine Actors

• Naver AI Overview API: https://apify.com/johnvc/naver-ai-overview-api?fpr=9n7kx3&fp_sid=skillrepo
• Brave AI Mode API: https://apify.com/johnvc/brave-ai-mode-api?fpr=9n7kx3&fp_sid=skillrepo
• Bing Copilot API: https://apify.com/johnvc/bing-copilot-api?fpr=9n7kx3&fp_sid=skillrepo
• Google News API: https://apify.com/johnvc/GoogleNewsAPI?fpr=9n7kx3&fp_sid=skillrepo