Bing Copilot API: How to Track Microsoft AI Answers and Citations in 2026

If your brand lives or dies on search, you now have a second problem: what the AI answer engines say about you. Bing Copilot answers millions of buying questions a day, and there is no report anywhere that tells you whether it mentions you or your competitor. I built the Bing Copilot API on Apify to close that gap: send a query, get Microsoft's AI answer back as structured JSON, citations and all.

Disclosure: the Apify links in this post are affiliate links. If you run the Actor, I may earn a referral commission at no extra cost to you.

Is there a Bing Copilot API?

Microsoft ships plenty of developer products around Copilot, but none of them return the consumer Copilot answer for a query the way a user sees it, with its cited sources attached. For AEO (answer engine optimization) work, that user-facing answer is the only artifact that matters: it is what your prospects actually read. So a Bing Copilot API in practice means a scraper you call like an API: one query in, one captured answer out.

What the Bing Copilot API returns

The Bing Copilot API returns Microsoft's live AI answer for a query as structured JSON: the answer header, the full answer as markdown, the raw text blocks, every cited reference, and an optional brand-mention flag.

Field	Example	Notes
answer header	"Top CRM picks for startups"	Headline summary of the answer
markdown	full answer document	One markdown doc, ready to store and diff
text blocks	paragraphs, headings, lists, tables	The answer's raw structure
references	title, link, snippet, source	Every citation Copilot showed
brandMentioned	true	Whether your tracked brand appears in the answer or its citations
summary	one line	Agent-ready recap of the answer

Featured media links come along when Copilot includes them.

Who this is for

SEO and AEO teams who need to know whether the brand shows up in AI answers, content strategists working out which domains Copilot likes to cite for a topic, and developers building their own AI visibility tooling who want raw answer data instead of someone else's dashboard.

The manual way, and where it breaks

The manual version is typing queries into Copilot and pasting answers into a spreadsheet, which dies at about query ten. The automated DIY version is worse than it looks: you are scripting a chat interface that streams its output, holds session state, and changes markup without notice. Extracting the citations cleanly, the part you actually need for AEO, is the fiddliest bit. I speak from a weekend I want back.

The faster way: run the Bing Copilot API

Apify Console

1. Open the Bing Copilot API and click Try for free.
2. Enter a query (or a list in queries) and, optionally, a brandToTrack.
3. Run it and download the dataset as JSON, CSV, or Excel.

REST

curl -X POST "https://api.apify.com/v2/acts/johnvc~bing-copilot-api/runs?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "query": "best crm for startups", "brandToTrack": "Salesforce" }'

Run endpoint reference: the Apify API docs. The task Bing Copilot answers API is this call saved and ready to run.

Check brand mentions in Python

import json
from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run = client.actor("johnvc/bing-copilot-api").call(
    run_input={
        "queries": ["best crm for startups", "best help desk software"],
        "brandToTrack": "Salesforce",
    }
)

for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    print("Salesforce mentioned:", row["brandMentioned"])
    print(json.dumps(row, indent=2)[:500])

Each query returns one answer row; the brandMentioned boolean gives you the headline metric, and the markdown plus references carry everything you need for the deeper analysis.

Track the whole answer-engine field, not one engine

Copilot is one of four engines I track the same way, next to Google AI Overview, Naver AI Overview, and Brave AI Mode, so you can compare how each one answers the same query set. The task AEO suite AI answer tracking shows the multi-engine setup.

AEO tracking on Bing Copilot

For a straight Copilot monitoring loop, AEO tracking with Bing Copilot runs a target query set and keeps the structured answers for comparison over time.

AI citation tracking

The references array tells you which domains Microsoft's answer engine trusts on your topic, which is the closest thing AEO has to a ranking factor. AI citation tracking with Bing Copilot pulls those citations out per query.

Brand monitoring in Microsoft's answer engine

Set brandToTrack to your brand, run your buying-intent queries, and read the booleans. Brand monitoring with Bing Copilot is that workflow saved as a task.

Use it from Claude and other MCP clients

The Actor is MCP-ready, so Claude, Claude Code, and Cursor can call it as a tool mid-conversation: ask "does Copilot mention us for our top ten queries" and the agent runs the checks and reads the results back. You can read more about Claude and Claude Code at claude.ai.

FAQ about scraping Bing Copilot

Why use an AI answer scraper instead of an AEO dashboard?

Dashboard platforms sell polished multi-engine reports on their schedule and their format. A scraper gives you the raw answer data as JSON you own, so you can diff answers, join them with your analytics, and build the exact report your team wants. If you are a developer, the data API is the better primitive.

How much does the Bing Copilot scraper cost?

It bills per query: one query resolved, one event charged, whether you send a single question or a list. New Apify accounts include free platform credit, so a first brand audit usually costs nothing out of pocket.

Can the scraper tell me which sources Copilot cites?

Yes, that is half the point. Every answer row includes the cited references with title, link, snippet, and source metadata, so you can rank the domains Copilot leans on for your topic and target your content or PR there.

Can Claude call this scraper through MCP?

Yes. Connect the Apify MCP server to Claude, Claude Code, or Cursor and the Actor shows up as a callable tool for live answer checks inside a conversation.

Can I schedule the scraper to track answer drift?

Yes, and drift is the reason to bother. Save your query set as a task, attach an Apify schedule, and store the markdown output from each run; diffing those documents shows exactly when Copilot's answer about you changed. Start from the Bing Copilot API.

What are the limits of this Bing Copilot scraper?

It covers one engine, Bing Copilot, with one answer per query, and generation takes a few seconds per query because the upstream engine is generating text. Answers are generative, so expect natural variation between runs; that is a property of the medium, and it is why trend tracking beats single snapshots.

More from Truffle Pig Data

The sibling engines in the AEO suite return the same kind of structured answers: the Google AI Overview API for Google's AI answers, the Naver AI Overview API for the Korean market, and the Brave AI Mode API for Brave's answer engine.

Wrapping up

You cannot optimize for an answer engine you cannot see. The Bing Copilot API turns Microsoft's AI answers into JSON you can store, diff, and act on.