Stack Overflow & Stack Exchange Q&A Scraper, No Login

Stack Overflow & Stack Exchange Q&A Scraper exports questions with their answers from Stack Overflow and the other Stack Exchange sites (Super User, Server Fault, Ask Ubuntu, Mathematics and 170+ more). You pick a site and a tag, a search query or a list of question links, and get one table row per question with the title, body, tags, votes, author and the answers inside the same row.

Use it to follow questions about your product or library, to build a Q&A dataset for search or an LLM, or to save a fixed list of threads. The smallest run is one tag, for example apify, and 10 questions. No login or API key is needed: the Actor reads the official Stack Exchange API, and every row has a ready CC BY-SA attribution line.

Table of contents

• What it does
• How it works
• Quick start
• Use cases
• Input
• Output
• Pricing
• Limits
• API, export and AI agents
• Troubleshooting
• FAQ
• Related Actors
• Support

What it does

The Actor reads questions from one Stack Exchange site and saves each question as one dataset row with its answers inside. It has three modes:

mode	What you give	What it returns
tag	site and up to 5 tags	Questions that carry all the tags, in your sort order
search	site, a query and optional tags	Questions that match the words in the title or body
question	questions: IDs or question URLs	Exactly these questions; a URL sets its own site

With includeAnswers on (the default), the answers are inside the row: the accepted answer first, then by score. You choose the body format: Markdown as written, HTML as shown on the site, or plain text. You can also limit the creation date with fromDate and toDate.

Good fit

• Watch new questions about your product tag on Stack Overflow.
• Build a question-and-answer dataset in Markdown for search, RAG or an LLM.
• Save a list of known threads with all answers and votes.
• Find the most-voted questions on a topic on any Stack Exchange site.

👉 Try it now: open the input form

How it works

1. The Actor asks the official Stack Exchange API for the questions of your tag list, search or question links on the site you chose.
2. For each page of questions it reads the answers, the content licence and the authors, sorts the answers (accepted first, then by votes) and converts the bodies to the format you picked.
3. It saves one row per question in the dataset, free not_found and error rows for links that do not work, and a RUN_SUMMARY record with counts and the remaining daily API quota.

👉 Try it now: open the input form

Quick start

1. Open the Input tab. Mode stays at the default Questions by tag (box 1).

2. Keep Site at the prefilled stackoverflow and Tags at the prefilled apify, or type your own site and tag (box 2).

3. Keep Max questions at the prefilled 10 or set your own number (box 3; the runtime default is 50).

4. Click Start at the bottom of the page.

5. Open the Questions view in the Output tab and check title, score, answerCount and link (box 5).

   

Expected result: the 10 newest questions with the apify tag and their answers, in a few seconds.

👉 Try it now: open the input form

Use cases

Buyer job	Input	Output	Key fields	Next step
Follow questions about your product	tags: ["your-tag"], sort: newest, scheduled	question rows	title, link, createdAt, isAnswered	Send unanswered ones to your support channel
Build an LLM or RAG dataset	tags, sort: votes, bodyFormat: markdown	rows with answers	body, answers[].body, answers[].isAccepted, attribution	Load into a vector store, keep the credit line
Research a topic	mode: search, query, fromDate	question rows	score, viewCount, tags	Count questions per month or per tag
Save known threads	mode: question, questions	one row per thread	answers, acceptedAnswerId, lastActivityAt	Archive or compare over time

👉 Try it now: open the input form

Input

You need a site and at least one of: tags (tag mode), query or tags (search mode), questions (question mode). A wrong input stops the run with INVALID_INPUT before any request.

Field	Plain name	Type	Default	Allowed values	What it changes	Recommendation
mode	Mode	enum	tag	tag, search, question	Which list the Actor reads	tag to follow a topic, question for known threads
site	Site	string	stackoverflow (prefill too)	API site name or domain: superuser, serverfault, askubuntu, math, math.stackexchange.com …	The Stack Exchange site	Use the short name from the site URL
tags	Tags	array of strings	none (prefill ["apify"])	up to 5 tags	Questions must carry all the tags. Empty in tag mode = all questions of the site	One or two tags give the clearest list
query	Search query	string	none	up to 500 characters	Words to find in titles and bodies (search mode only)	Add sort: relevance
questions	Question IDs or URLs	array of strings	none	up to 1,000 IDs like 11227809 or question URLs	Exact questions to read (question mode only). Duplicates are read once	Paste links from your browser
sort	Sort	enum	newest	newest, active, votes, hot (tag mode), relevance (search mode)	Order of the list	votes for the best-known questions
maxQuestions	Max questions	integer	50 (prefill 10)	1–1,000	The run stops after this many question rows	Start small, then raise
includeAnswers	Include answers	boolean	true	true, false	Answers inside each row	Turn off for a list of titles only
maxAnswersPerQuestion	Max answers per question	integer	30	1–500	Keep only the best answers	answersComplete shows if all answers fit
bodyFormat	Body format	enum	markdown	markdown, html, text	Format of question and answer bodies	markdown for LLMs, text for spreadsheets
fromDate, toDate	Created from / until	date	none	2025-01-01 (UTC; toDate includes the whole day)	Only questions created in this window	Use for a monthly report
apiKey	Stack Exchange API key	secret string	none	free key from stackapps.com	Higher daily quota	Not needed for normal runs

Interacting settings

When	Field A	Field B	Effect
Search	query	tags	Both must match: the words and all the tags
Order	sort: hot	mode: tag	hot works only in tag mode; relevance only in search mode
Question links	questions with a URL	site	The URL's own site wins; bare IDs use site
Answers	includeAnswers: false	maxAnswersPerQuestion	No answers, so maxAnswersPerQuestion is ignored and answers is null
Dates	fromDate, toDate	mode: question	Dates work in tag and search modes only

Input examples

Newest questions about a product tag (the prefill)

{
  "site": "stackoverflow",
  "tags": ["apify"],
  "maxQuestions": 10
}

Expected: 10 question rows, newest first, each with its answers. The output example below comes from this input.

Best-voted answers on Server Fault for a search

{
  "mode": "search",
  "site": "serverfault",
  "query": "nginx reverse proxy",
  "sort": "votes",
  "maxQuestions": 100,
  "bodyFormat": "text"
}

Expected: up to 100 question rows, most votes first, with plain-text bodies and answers.

A fixed list of threads

{
  "mode": "question",
  "questions": [
    "11227809",
    "https://serverfault.com/questions/997924"
  ],
  "maxAnswersPerQuestion": 5
}

Expected: one row per question that exists, each with up to 5 answers. A missing question gives a free not_found row.

Coming from another Stack Exchange Actor

Input names here are different. This table helps to move an existing input:

Common input in other Actors	This Actor
tagged: "a;b"	tags: ["a", "b"]
searchQuery	mode: "search", query
questionIds	mode: "question", questions
fetchAnswers	includeAnswers (on by default, no extra price)
maxResults	maxQuestions
sort: activity / creation	sort: active / newest

Output fields are camelCase (questionId, answerCount) with ISO dates, not raw API snake_case. The asker is one author object.

👉 Try it now: open the input form

Output

Output example

A real row from the prefill run (body and the answer body are shortened here):

{
  "site": "stackoverflow",
  "questionId": 77823456,
  "title": "Apify: weird costs for using key-value stores",
  "link": "https://stackoverflow.com/questions/77823456/apify-weird-costs-for-using-key-value-stores",
  "tags": ["node.js", "apify"],
  "score": 0,
  "upVotes": 1,
  "downVotes": 1,
  "viewCount": 213,
  "answerCount": 1,
  "isAnswered": true,
  "acceptedAnswerId": null,
  "commentCount": null,
  "bountyAmount": null,
  "isClosed": false,
  "closedReason": null,
  "closedAt": null,
  "createdAt": "2024-01-16T04:20:59.000Z",
  "lastActivityAt": "2024-01-16T13:03:25.000Z",
  "lastEditedAt": null,
  "author": {
    "userId": 2486638,
    "displayName": "Alvein",
    "reputation": 217,
    "userType": "registered",
    "profileUrl": "https://stackoverflow.com/users/2486638/alvein"
  },
  "bodyFormat": "markdown",
  "body": "I'm working with Node.js but I think my question is language-agnostic.\n\nIn my Apify actor, First I do a few long scrapes…",
  "answers": [
    {
      "answerId": 77825927,
      "link": "https://stackoverflow.com/questions/77823456/apify-weird-costs-for-using-key-value-stores/77825927#77825927",
      "score": 1,
      "upVotes": 1,
      "downVotes": 0,
      "isAccepted": false,
      "commentCount": 1,
      "createdAt": "2024-01-16T13:03:25.000Z",
      "lastActivityAt": "2024-01-16T13:03:25.000Z",
      "lastEditedAt": null,
      "author": {
        "userId": 9898465,
        "displayName": "Lukáš Křivka",
        "reputation": 983,
        "userType": "registered",
        "profileUrl": "https://stackoverflow.com/users/9898465/luk%c3%a1%c5%a1-k%c5%99ivka"
      },
      "body": "Apify doesn't charge for how big the writes are. Apify charges for number of writes and then for time based storage…",
      "license": "CC BY-SA 4.0",
      "licenseUrl": "https://creativecommons.org/licenses/by-sa/4.0/"
    }
  ],
  "answersIncluded": 1,
  "answersComplete": true,
  "license": "CC BY-SA 4.0",
  "licenseUrl": "https://creativecommons.org/licenses/by-sa/4.0/",
  "attribution": "\"Apify: weird costs for using key-value stores\" by Alvein (https://stackoverflow.com/users/2486638/alvein), https://stackoverflow.com/questions/77823456/apify-weird-costs-for-using-key-value-stores, licensed under CC BY-SA 4.0",
  "rank": 8,
  "input": null,
  "inputs": null,
  "status": "complete",
  "statusReason": null,
  "errorCode": null,
  "sourceUrl": "https://api.stackexchange.com/2.3/questions?site=stackoverflow&order=desc&sort=creation&pagesize=10&filter=…&tagged=apify&page=1",
  "fetchedAt": "2026-10-17T12:18:30.349Z"
}

null means the API gave no value (no accepted answer, no bounty, never edited). 0 is a real zero. commentCount can be null in tag and search lists and filled in question mode: the Actor copies what the API endpoint returns.

Output fields

Question rows (status: complete or partial)

Field	Meaning	Type	Can be null?	Example
site	Site the row came from	string	no	"stackoverflow"
questionId	Question ID, stable key	integer	no	77823456
title, link	Title and public question link	string	no
tags	Question tags	array	no	["node.js", "apify"]
score, upVotes, downVotes	Votes	integer	yes	0, 1, 1
viewCount, answerCount	Views and number of answers	integer	yes	213, 1
isAnswered, acceptedAnswerId	Has an upvoted or accepted answer; ID of the accepted one	boolean, integer	yes	true, null
commentCount, bountyAmount	Comments on the question; open bounty	integer	yes	null
isClosed, closedReason, closedAt	Closed questions	boolean, string	yes	false
createdAt, lastActivityAt, lastEditedAt	Dates (ISO, UTC)	string	yes	"2024-01-16T04:20:59.000Z"
author	userId, displayName, reputation, userType, profileUrl	object	yes (deleted users)
bodyFormat, body	Body format and question body	string	body yes	"markdown"
answers	Answers: answerId, link, score, isAccepted, author, body, license, dates	array	null when includeAnswers is off
answersIncluded, answersComplete	How many answers are in the row; false if maxAnswersPerQuestion cut some	integer, boolean	yes	1, true
license, licenseUrl	Content licence of the question	string	yes	"CC BY-SA 4.0"
attribution	Ready credit line: title, author, link, licence	string	yes
rank	Position in the tag or search list	integer	yes (question mode)	8
status, statusReason, errorCode	Row status and why	string	statusReason, errorCode yes	"complete"
input, inputs	What you entered for this row (question mode)	string, array	yes	"11227809"
sourceUrl, fetchedAt	API request and fetch time	string	no

partial means the row has no licence from the API, or the API returned fewer answers than expected. It is still a useful row and is charged.

Issue rows (status: not_found or error)

Question mode only. These rows are free.

• not_found: the IDs do not exist, are deleted or are not question IDs. Missing IDs from one batch share one row, with the list in inputs.
• error: a bad item (for example an answer link or a non-Stack Exchange URL), or a source error for that batch. Items with the same reason share one row, with the list in inputs.

When a row is about one item, input and questionId are filled; when it is about several, they are null and inputs lists them.

Run report

The key-value store record RUN_SUMMARY has the run status, reason, counts (rows, questions, answers, complete, partial, not found, errors, charged questions), request counts and the remaining daily API quota. It never contains your API key.

Run status	Meaning	Run result
COMPLETE	All requested questions were read	SUCCEEDED
CAPPED	Stopped early with saved rows: your maximum charge (CHARGE_LIMIT), the daily API quota (QUOTA_LIMIT), the time limit (DEADLINE) or the API page limit (PAGE_LIMIT)	SUCCEEDED, saved rows are kept
NO_RESULTS	No questions match the filters	SUCCEEDED, empty dataset
NOT_FOUND	No question from the list exists	SUCCEEDED, free not_found rows
PARTIAL	Some rows are partial (PARTIAL_ROWS), or the source failed after some rows (SOURCE_ERROR)	SOURCE_ERROR marks the run FAILED; saved rows are kept
ERROR	The source failed or the run stopped before any question was saved	FAILED
INVALID_INPUT	The input cannot work, for example an unknown site, or no item in questions is a valid question	FAILED
CANCELLED	You aborted the run	saved rows are kept

Licence and attribution

Stack Exchange content is licensed by its authors under CC BY-SA (2.5, 3.0 or 4.0, by post date). Each row has its license and licenseUrl, and attribution gives a ready credit line with the title, author and link. When you reuse the content, show that line and keep the same licence. The Stack Exchange API terms also ask you to name Stack Exchange as the source. The Stack Exchange Acceptable Use Policy limits some uses, for example training AI models without their consent; check that it fits your plan.

Data comes from the Stack Exchange API. See also the Stack Exchange API terms of use, the Acceptable Use Policy and the CC BY-SA 4.0 licence. This Actor is not affiliated with or endorsed by Stack Exchange, Inc.

👉 Try it now: open the input form

Pricing

Pay per event. Apify platform usage is included.

Event	When	Price (USD)
Actor start	once per run; Apify counts the start per GB of run memory, so it is one start at the default 256 MB and the 512 MB maximum	0.002
Question	per saved question row (complete or partial), answers included	0.0027 Free · 0.00243 Bronze · 0.00216 Silver · 0.00189 Gold and above

not_found and error rows are free. A run with no results pays only the start.

Example run	Free	Bronze	Silver	Gold+
1 question	$0.0047	$0.0044	$0.0042	$0.0039
10 questions (the prefill)	$0.029	$0.026	$0.024	$0.021
100 questions	$0.272	$0.245	$0.218	$0.191
1,000 questions with answers	$2.702	$2.432	$2.162	$1.892
Search with no matches	$0.002	$0.002	$0.002	$0.002

Set Maximum cost per run in the run options to cap a run. The Actor checks the limit before each row. It stops before a charge would go above it, keeps what it already saved and ends as CAPPED with reason CHARGE_LIMIT.

👉 Try it now: open the input form

Limits

• Up to 1,000 questions per run. Measured runs of 1,000 questions with answers, all COMPLETE inside the 600-second default timeout: 16 seconds for Super User windows-10 in HTML, 217 seconds for the most-voted Stack Overflow javascript questions in Markdown (about 20 answers per question, 40 MB of text).
• Without an API key, the Stack Exchange API gives about 300 requests per day per IP address, and Apify runs share IP addresses. One request reads up to 100 questions with answers, so normal runs use a few requests. RUN_SUMMARY shows the remaining quota.
• If the daily quota runs out, the run stops as CAPPED with reason QUOTA_LIMIT, keeps saved rows and suggests an apiKey. If nothing was saved yet, the run is ERROR.
• The API asks clients to wait between some requests. The Actor follows this, so big runs take minutes, not seconds.
• The Actor stops reading 30 seconds before the run timeout, and after 870 seconds at most, so it can save the results (CAPPED, DEADLINE). With a shorter timeout, for example 300 seconds, a run of 1,000 answer-heavy questions can stop early with the rows read so far.
• No comments, user profiles, badges or chat. Only questions and their answers.
• No private or deleted content. Deleted questions come back as not_found rows.
• Answer links (/a/<id>) are not looked up. Use the question link instead.
• The data is live: votes, views and answers can change between runs.

👉 Try it now: open the input form

API, export and AI agents

• Export: the dataset downloads as JSON, CSV, Excel, XML or HTML. The Questions view shows a flat table with the main question fields.
• API: start a run with POST https://api.apify.com/v2/acts/cliqtomedia~stack-exchange-scraper/runs and the input JSON; read rows from the run's default dataset.
• Scheduling, webhooks, Make, Zapier, n8n: use the standard Apify integrations. The next system can use site + questionId as a stable key.

Use with AI agents (MCP)

• MCP server with only this Actor: https://mcp.apify.com?tools=cliqtomedia/stack-exchange-scraper
• Smallest useful input: {"site": "stackoverflow", "tags": ["apify"], "maxQuestions": 10}
• Key output fields: title (question title), link (public URL), score (votes), answers[].body with answers[].isAccepted (answer text and accepted flag), attribution (credit line to show with the content).
• The prefill run takes a few seconds and costs $0.029 on the Free plan (start $0.002 + 10 questions). 1,000 questions with answers took 16 to 217 seconds in our tests, inside the 600-second default timeout. Keep maxQuestions small for quick answers.

👉 Try it now: open the input form

Troubleshooting

The run failed with Unknown Stack Exchange site

Symptom: the run is FAILED with INVALID_INPUT. Likely reason: the site value is not a site name. Check: open the site in a browser; the name is the first part of the address (superuser.com → superuser, math.stackexchange.com → math). Action: fix site or paste the full domain.

The dataset is empty

Symptom: the run SUCCEEDED with NO_RESULTS. Likely reason: no question has all the tags, or the dates or query are too narrow. Check: open the tag page on the site. Action: remove a tag or widen the dates.

I got a not_found row

Likely reason: the question was deleted, or the ID is an answer ID. Check: open the link from inputs in a browser. Action: use the question link. These rows are free.

The run ended with QUOTA_LIMIT

The shared daily API quota of this IP address ran out. Saved rows are kept. Run again later, or add a free Stack Exchange API key in apiKey.

👉 Try it now: open the input form

FAQ

Do I need a Stack Overflow account or API key?

No. The Actor reads public data from the official Stack Exchange API. No login and no cookies. The apiKey field is optional: a free Stack Apps key only raises the daily quota and gives no account access.

Are answers charged separately?

No. Answers are part of the question row. One question row is one charge, with up to 500 answers inside.

Which sites work?

Any public Stack Exchange site: Stack Overflow, Super User, Server Fault, Ask Ubuntu, Mathematics, MathOverflow, Stack Overflow in Russian and others. Use the site name or domain in site.

Are the numbers live?

Yes. Each run reads the API at that moment; fetchedAt shows when. Votes, views and answers change over time.

Is personal data included?

Only public profile data shown on the site: the author's user ID, display name, reputation and profile link. No email, location or other profile details.

👉 Try it now: open the input form

Related Actors

• Hacker News Scraper: Stories, Comments & Parent IDs — stories and comment threads from another big developer community.
• Reddit Scraper: Posts, Comments & Subreddits (No login) — posts and comments from subreddits about the same topic.
• Quora Answers Scraper: Upvotes, Views & Authors, No Login — questions and answers from Quora, with upvotes and authors.
• GitHub Releases Scraper & Monitor: Notes & Assets — release notes of the library your questions are about.

👉 Try it now: open the input form

Support

For a bug, question or feature request, open an issue on the Actor's Issues tab. Add the run ID and your input. Never post API keys or other secrets.

👉 Try it now: open the input form