ChatGPT API
Send the prompt your buyers type and get back what ChatGPT tells them: the full answer, every page it cites, the searches it ran and the products it recommends. AI visibility data as one POST.
50 free credits on signup, enough for 25 prompts. No card required.
{ "data": { "platform": "chatgpt", "prompt": "What are the best running shoes for beginners?", "model": "gpt-6", "answer": { "markdown": "The best running shoes for beginners are usually **comfortable, well-fitting daily trainers** that feel good from your very first run...\n\n## 1\\. Brooks Ghost 18\n\n(Best overall for beginners)..." }, "sources": [ { "url": "https://www.rtings.com/running-shoes/reviews/best/beginners", "title": "The 6 Best Running Shoes for Beginners of 2026", "snippet": null, "source": "RTINGS.com", "published_at": "2026-09-18T00:00:00.000Z" }, { "url": "https://www.runnersworld.com/gear/a69991542/editors-beginner-shoe-recommendations/", "title": "Runner’s World Editor Shoe Recommendations for Beginners", "snippet": null, "source": "Runner's World", "published_at": "2026-01-21T00:00:00.000Z" } ], "search_queries": [ "best running shoes for beginners 2026 review Brooks Ghost ASICS Nimbus Saucony Ride Nike Pegasus" ], "shopping": [ { "title": "Brooks Ghost 18", "url": null, "price": null, "image_url": null, "merchant": null }, { "title": "ASICS Novablast 6", "url": null, "price": null, "image_url": null, "merchant": null }, { "title": "HOKA Clifton 11", "url": null, "price": null, "image_url": null, "merchant": null } ] }, "response_time": 28081, "credits_used": 2, "credits_remaining": 4998}Live call against the real endpoint. Answers take 10-40 seconds. 2 free runs, then it's your own key.
What is the Scavio ChatGPT API?
The Scavio ChatGPT API is a REST API that sends a prompt to ChatGPT and returns its answer as structured JSON: the answer as Markdown, the pages it cited with title, publisher and date, the web searches it ran, and the products it recommended. It is built for tracking brand mentions and AI visibility.
- 1
- endpoint, live in production
- 2
- credits per answered prompt
- 0
- credits for a failed prompt
One endpoint, the whole answer
What ChatGPT says, who it credits, and what it recommends.
Prompt
POST /api/v1/chatgpt/promptSend the question your buyer would ask. Get back the answer ChatGPT gives, every page it cites with title, publisher and date, the web searches it ran to answer, and the products it put in front of the reader.
The ChatGPT endpoint
A POST that returns JSON and takes the same API key as every other Scavio platform. 2 credits per answered prompt.
| Endpoint | Returns | Credits |
|---|---|---|
Prompt/api/v1/chatgpt/prompt | ChatGPT's answer as Markdown, the pages it cites, the searches it ran and the products it recommended | 5 |
50 free credits on signup covers 25 prompts. See plans
Is your brand in the answer?
Check the answer, the sources and the recommended products for your brand. The response on the right is a real capture, trimmed.
from scavio import ScavioClient
client = ScavioClient(api_key="sk_live_your_key")
res = client.chatgpt.prompt("What are the best running shoes for beginners?")
answer = res["data"]
# Is your brand in the answer?
print("Brooks" in answer["answer"]["markdown"])
# Which pages did ChatGPT cite?
for s in answer["sources"]:
print(s["source"], s["url"])
# Which products did it recommend?
print([p["title"] for p in answer["shopping"]]){
"data": {
"platform": "chatgpt",
"prompt": "What are the best running shoes for beginners?",
"model": "gpt-6",
"answer": {
"markdown": "The best running shoes for beginners are usually **comfortable, well-fitting daily trainers** that feel good from your very first run...\n\n## 1\\. Brooks Ghost 18\n\n(Best overall for beginners)..."
},
"sources": [
{
"url": "https://www.rtings.com/running-shoes/reviews/best/beginners",
"title": "The 6 Best Running Shoes for Beginners of 2026",
"snippet": null,
"source": "RTINGS.com",
"published_at": "2026-09-18T00:00:00.000Z"
},
{
"url": "https://www.runnersworld.com/gear/a69991542/editors-beginner-shoe-recommendations/",
"title": "Runner’s World Editor Shoe Recommendations for Beginners",
"snippet": null,
"source": "Runner's World",
"published_at": "2026-01-21T00:00:00.000Z"
}
],
"search_queries": [
"best running shoes for beginners 2026 review Brooks Ghost ASICS Nimbus Saucony Ride Nike Pegasus"
],
"shopping": [
{ "title": "Brooks Ghost 18", "url": null, "price": null, "image_url": null, "merchant": null },
{ "title": "ASICS Novablast 6", "url": null, "price": null, "image_url": null, "merchant": null },
{ "title": "HOKA Clifton 11", "url": null, "price": null, "image_url": null, "merchant": null }
]
},
"response_time": 28081,
"credits_used": 2,
"credits_remaining": 4998
}Why not the OpenAI API?
The OpenAI API answers your request. AI visibility is about the answer ChatGPT gives everyone else.
OpenAI API
- Returns a completion from the model and tools you configure, not the answer ChatGPT gives its users
- Sources and searches only when you wire up search tools yourself, and they are your choices, not ChatGPT's
- No product recommendations as ChatGPT shows them to shoppers
- Billed per token, so a long answer costs more than a short one
Scavio ChatGPT API
- The answer ChatGPT gives, with the sources it chose to cite
- The web searches it ran and the products it recommended, as arrays
- A flat 2 credits per answered prompt; failed prompts are free
- Same key and JSON envelope as Gemini and 40+ other Scavio platforms
What teams build with it
AI visibility tracking
Run the prompts your buyers type, on a schedule, and record whether your brand is named in the answer. That share of mentions is the number the GEO dashboards sell; this is the raw data behind it.
promptCitation and source analysis
Every answer comes with the pages ChatGPT cited, with title, publisher and publish date. Find which review sites and listicles decide your category, then earn a place on them.
promptProduct recommendation monitoring
For commercial prompts ChatGPT puts products in front of the reader. shopping[] lists them in order, so you can track whether your product is recommended and which competitors outrank it.
promptQuery fan-out research
search_queries[] is what ChatGPT actually searched to answer. Those are the queries your content has to win, and they are rarely the prompt as typed.
promptCross-assistant comparison
Send the same prompt to the Gemini endpoint and compare. One response shape for both, so one parser covers both assistants.
chatgpt + geminiHow to get ChatGPT answers as JSON
- 1
Get an API key
Sign up for a Scavio account and copy your key from the dashboard. You get 50 credits free on signup, enough for 25 prompts, with no card required.
- 2
POST a prompt
Send {"prompt": "What are the best running shoes for beginners?"} to https://api.scavio.dev/api/v1/chatgpt/prompt with an Authorization: Bearer header. Add country to localize the answer.
- 3
Read the answer, sources and products
data.answer.markdown is the full answer. data.sources lists the pages it cited, data.search_queries the searches it ran, and data.shopping the products it recommended. Check them for your brand.
ChatGPT API FAQ
What is the Scavio ChatGPT API?
+
It is a REST API that sends a prompt to ChatGPT and returns the answer ChatGPT gives as structured JSON: the full answer as Markdown, the pages it cited, the web searches it ran and the products it recommended. One POST endpoint, authenticated with a Scavio API key.
How is this different from the OpenAI API?
+
The OpenAI API returns a completion from the model and tools you configure in your own request. It is not the answer ChatGPT gives the people asking it questions, with the sources ChatGPT chose and the products it showed. For AI visibility you need to see what your buyers see, and that is what this endpoint returns.
Who is this for?
+
SEO and GEO teams tracking brand mentions in ChatGPT, agencies reporting AI visibility for clients, ecommerce brands checking whether ChatGPT recommends their products, and developers building AI visibility tools who need the underlying data rather than a dashboard.
How do I track brand mentions in ChatGPT?
+
Build a list of the prompts your buyers ask, call the endpoint for each on a schedule, and record whether your brand appears in answer.markdown, sources and shopping. Over time that is your share of answer, and you can compare it with competitors on the same prompts.
Is this a ChatGPT scraper API?
+
It returns the same data a ChatGPT scraper would, without you running or maintaining one: no browsers, accounts or retries on your side. You send a prompt and get clean JSON back.
How much does a prompt cost?
+
2 credits per answered prompt. Failed prompts are not charged. New accounts get 50 free credits on signup, enough for 25 prompts. On the Growth yearly plan that works out to about $4 per 1,000 prompts.
How long does an answer take?
+
Typically 10 to 40 seconds, because ChatGPT writes the answer and often searches the web first. Set your HTTP client timeout to at least 90 seconds, and if a call returns 503, retry after the Retry-After header (it is never charged).
Are answers the same every time?
+
No. ChatGPT answers vary between runs, so treat visibility as a rate measured across several runs per prompt, not a single check. Responses are never cached, so every call is a fresh answer.
Can I choose the model or the country?
+
Yes. model defaults to auto, the model ChatGPT picks by default, and you can pass a specific model id. country takes a 2-letter code and localizes the answer. The model that actually answered comes back in the model field.
What does the response look like?
+
The standard Scavio envelope: data, response_time, credits_used and credits_remaining. data holds platform, prompt, model, answer.markdown, sources (url, title, snippet, source, published_at), search_queries and shopping (title, url, price, image_url, merchant). Set include_html to also get the answer as HTML.
Does it work with Gemini too?
+
Yes. The Gemini endpoint takes the same prompt and returns the same response shape, so you can track both assistants with one parser. See the Gemini API page. Gemini API
Which languages and tools are supported?
+
Official Python (pip install scavio) and JavaScript/TypeScript (npm i scavio) SDKs, a remote MCP server at mcp.scavio.dev for AI agents, and an n8n community node. Any language can call the REST endpoint with an HTTP client.
Explore more of Scavio
Gemini API
The same prompt, answered by Gemini, in the same shape.
Google AI Mode API
Google's generated answer and the sources it cited.
Google Search API
The full SERP, including AI Overview.
ChatGPT API docs
Parameters, response fields and errors.
MCP server
Every platform as a callable tool for agents.
Pricing
Credit packs and plans, no per-platform surcharge.
Find out what ChatGPT says about your brand
50 free credits on signup, no card. Your first request is a POST with a prompt in it.