AI News
13 Aug 2026
Read 8 min
How to query Cloudflare Radar API for instant charts
how to query Cloudflare Radar API to generate instant interactive charts for fast data-backed reports
Learn how to query Cloudflare Radar API for instant charts using Radar Researcher. You can type a plain question, get live data, and see interactive visualizations in seconds. No manual filters or query syntax. You can also start from any chart, ask “Explain with AI,” and keep digging with suggested follow-ups.
Cloudflare Radar has long shared open Internet data. Now Radar Researcher turns that data into fast answers. It reads your question, fetches the right endpoints, and shows accurate charts with short explanations. You can audit the steps, share a link, and return to the same conversation anytime.
Open radar.cloudflare.com and click the Researcher button in the header.
Ask a question like “What is Internet quality in Portugal this month?”
See interactive charts plus a short summary, with suggested next questions.
Choose a brief answer or a fuller report when prompted.
On any Radar visualization, click “Explain with AI.”
Researcher reads the chart screenshot, the exact data behind it, and your filters.
It explains what you are seeing, using the precise numbers from Radar’s API.
Researcher skips guesswork. It searches the OpenAPI spec, calls the right endpoint, and pulls live results.
It returns a lightweight chart spec that the site renders with Radar’s own components, so numbers stay exact.
You can expand a trace to see which datasets it used and how it reasoned.
If you prefer code, you still get the benefit of the model’s search and execution steps. This is still how to query Cloudflare Radar API without hard-coding endpoints or memorizing parameters.
The model fetches data, then outputs a tiny spec that points to that exact API path.
Radar’s frontend matches the spec to the fetched data and renders the chart you know from Radar pages.
No rounding, no truncation, and no copy-paste errors from Markdown tables.
Time series and stacked areas for traffic or outages.
Maps, donuts, bar charts, histograms, and the Internet Quality “speed flower.”
Ask: “How good is home Internet in Portugal?”
Researcher pulls the Internet Quality dataset and shows speeds, latency, and consistency.
Follow up: “Compare Portugal with Spain” or “Show common outages in Portugal.”
Ask: “What happened to Internet traffic in Iran early 2026?”
Researcher finds Radar’s outage events and traffic index, then builds a clear timeline.
You get an annotated traffic chart and a table of outage periods. Next, compare with neighbors.
Runs on Cloudflare Workers with Durable Objects for chat state and history.
Uses Workers AI with a fallback chain of open models, so it keeps answering when one model is busy.
AI Gateway handles logging, caching, and safety.
Through Cloudflare’s MCP server and Code Mode, the model uses three tools (search, execute, docs) to reach any Radar endpoint.
No prompt hard-coding. As Radar adds datasets, Researcher can use them right away.
Small models write chat titles and suggest follow-ups without slowing your main reply.
Researcher knows the current date and high-level details of your connection, so “here” questions make sense.
Share conversations with a link that auto-expires after 30 days.
Use concise mode for a quick quote; switch to report mode for context and multiple views.
Bookmark key chats to build a living brief you can refresh as events change.
Launch from Radar’s search bar or use voice input when speed matters.
Audit the reasoning panel to confirm sources before you publish.
Fewer clicks: ask once, get the right chart.
Faster cycles: suggested follow-ups guide the next question.
Higher trust: every number comes from live endpoints you can verify.
Repeatable work: saved chats act like recipes you can run again.
Agent-ready web: with WebMCP tools, browser agents can drive Radar features without scraping.
Cloudflare Radar Researcher makes live Internet data easy to explore, explain, and share. If you want a faster path on how to query Cloudflare Radar API and turn answers into instant charts, open the Researcher panel on radar.cloudflare.com and start asking questions.
How to query Cloudflare Radar API in plain language
Start a chat from anywhere on Radar
Turn any chart into a conversation
Why this feels instant
Build instant charts without manual parsing
How charts stay accurate
What you can visualize
Practical examples you can try
Check Internet quality in Portugal
Investigate an Internet disruption
Power under the hood, kept simple for you
Reliability and speed
Flexible data access
Helpful extras
Tips for teams, reporters, and operators
Why this changes how to query Cloudflare Radar API
(Source: https://blog.cloudflare.com/introducing-radar-researcher/)
For more news: Click Here
FAQ
Q: What is Radar Researcher and what does it do?
A: Radar Researcher is a beta AI tool on Cloudflare Radar that answers plain-language questions with real, interactive charts built from Radar’s API. It fetches live data, renders the same visualizations used across Radar, and provides short explanations.
Q: How do I start a chat with Radar Researcher on radar.cloudflare.com?
A: Open radar.cloudflare.com and click the Researcher button in the header to dock the panel beside the page so you can browse while you chat. Ask a plain question like “What is Internet quality in Portugal this month?” and get interactive charts plus a short summary.
Q: Can I turn any existing Radar chart into an explanation?
A: Yes — click “Explain with AI” on any Radar visualization to open Researcher with that exact chart, and it will read the screenshot, the exact API data, and your current filters. Researcher then analyzes the data and explains what you are seeing using the precise numbers from Radar’s API.
Q: How does Radar Researcher find the right endpoints and produce accurate charts?
A: Researcher searches Radar’s OpenAPI spec, writes and executes small code snippets to call the correct endpoints, and returns a lightweight chart spec that points to the fetched API path so Radar’s frontend renders the exact visualization. This is essentially how to query Cloudflare Radar API without hard-coding endpoints or memorizing parameters.
Q: Can I audit Researcher’s reasoning and verify the data it used?
A: Yes — every answer includes an expandable trace that shows which datasets were queried and the exact tool calls the model made, so you can see how it reached its answer. You can expand the trace to view the API paths the assistant used and verify numbers against the live endpoints.
Q: What technology powers Radar Researcher and how are sessions persisted?
A: Radar Researcher runs on Cloudflare Workers with the Cloudflare Agents SDK, and each conversation is a stateful Durable Object with a SQLite database so chat history and streaming responses persist. It uses Workers AI with an ordered fallback across open models and routes calls through AI Gateway for logging, caching, and safety.
Q: Can I save or share conversations and how long do shared links last?
A: Conversations are saved in a searchable history you can pin and revisit, and you can share any conversation with a link that automatically expires after 30 days. Shared conversations are stored via R2 and the system applies per-IP rate limiting.
Q: How does Radar support browser agents or automation with Researcher?
A: Radar implements WebMCP support for both the imperative and declarative APIs so browser agents can discover and call tools instead of scraping the DOM, letting agents filter by location, change date ranges, search entities, or drive existing UI features. This makes it easier for automation to interact with Radar without reverse-engineering the site.
Contents
