The reliable Reddit scraper API

The Reddit scraper that keeps working when Reddit changes.

UGC Scraper routes every request through automatic fallback, so when Reddit reprices the API, kills an endpoint, or blocks an IP, you still get the same clean JSON. Scrape any Reddit post, subreddit, user, or search, and get back one consistent schema. No Reddit API keys, no proxies to babysit, no 3am pages.

Get your API key See how it works

No Reddit credentials · Free tier · Live in 2 minutes

POST /v1/scrape

# one call, any Reddit post
curl -X POST https://api.ugcscraper.com/v1/scrape \
  -H "Authorization: Bearer rps_live_…" \
  -d '{"url":"reddit.com/r/…"}'

# → one consistent JSON schema
{
  "title": "Best note-taking app?",
  "author": "reddit_user",
  "score": 1240,
  "num_comments": 318,
  "comments": [
    { "author": "dev", "score": 92 }
  ]
}

See the shape

Try a scrape, right here

A real scrape of a live Reddit thread. Run it to see the JSON and the CSV export.

A real captured scrape. Sign up for a key to scrape any URL.

Why UGC Scraper

Built to survive Reddit's next change, not just today's

Pushshift died. The API got repriced. The .json trick got killed. Every one of those broke thousands of pipelines overnight. UGC Scraper runs automatic fallback behind one endpoint, so when one path gets blocked or deprecated, another takes over before you'd ever notice. Your integration never changes. Your data never stops.

Failover you never have to trigger

A blocked or rate-limited request fails over automatically, mid-request, with no retry logic on your side.

The same JSON, no matter how it was fetched

Title, author, score, comments: one shape, every time. Build once; it never shifts under you, no matter how the data was fetched.

Your data won't disappear (like the last tool did)

Every scrape is an immutable snapshot you can re-pull forever. No rug-pulls, no deleted archives.

Re-run analyses without re-scraping

Repeat pulls return from a shared cache in milliseconds, no new request, no new charge.

Predictable by design

Per-key quotas and burst limits out of the box. You always know your ceiling, no surprise spikes.

See we're up, in the open

A public status page shows live operational status, extraction success rate, and typical latency. Every scrape also reports its method, cache status, and latency in the response headers.

One API, one schema

Scrape Reddit posts, comments, users, subreddits, and search

Send a URL, name, or query and get back structured data, the same shape every time.

Posts

Title, body, author, score, timestamps, media, and flair.

Comments

The full thread with authors, scores, timestamps, and reply nesting, not just the top level. A real Reddit comment scraper, not a title grabber.

Users

A user's recent submissions as post summaries.

Subreddits

Recent posts from any subreddit, sorted new or top.

Search

Find posts by keyword, across Reddit or inside a subreddit.

Posts return a full RedditPost; subreddits, users, and search return post summaries. One consistent JSON shape across all of them.

One schema, every time

One consistent JSON schema, every time

Every request returns the exact same shape, no matter how it was fetched behind the scenes. Title, author, score, comment count, and the full comment tree, all in one schema your code can rely on. Build your parser once and it never shifts under you.

Prefer a spreadsheet? Export your history as CSV, one row per snapshot or one row per comment, in a single request.

POST /v1/scrape

// request
{ "url": "reddit.com/r/…" }

// response, same shape every time
{
  "title": "…",
  "author": "…",
  "score": 1240,
  "num_comments": 318,
  "comments": [
    { "author": "…", "score": 92, "body": "…" }
  ]
}

Three fewer things to break

No Reddit API keys, no proxies, no rate-limit bans

Most Reddit scrapers break on the same three things. UGC Scraper takes all three off your plate.

No Reddit API keys

No OAuth app, no developer approval, no Reddit credentials to manage. You sign up, get one UGC Scraper key, and start. Scrape Reddit without a Reddit API key at all.

No proxies to babysit

You never rent, rotate, or debug a proxy pool. The fallback layer handles the network for you.

No rate-limit bans

If a request is throttled or blocked, it fails over automatically, mid-request. Your IP never gets burned, and your pipeline does not stop.

The result is the thing nobody else promises: a Reddit scraper that keeps returning data when Reddit changes the rules.

Deeper than one listing

Get past Reddit's 1000-post listing limit

Reddit caps any single listing at about 1,000 items. Public endpoints, the official API, and most open-source scripts all hit the same wall, which is where a lot of scrapers quietly stop and leave you with one thin slice of a busy subreddit.

Comments are the part people usually care about, and they are not bound by that listing cap. On a large, heavily-ranked thread, a single request only surfaces one slice of the discussion. UGC Scraper pulls more of the thread than a single fetch returns, so you recover far more of the real comment tree.

Every fresh scrape is also written as a durable, immutable snapshot. You can re-pull any post you have already scraped straight from history without spending quota or hitting Reddit again, and because the snapshots are append-only you keep a real record of how a post's score and comments changed over time.

The honest version: we do not magically paginate a listing past Reddit's 1,000-item ceiling. What we do is recover the full comment tree per post, keep everything you pull, and route around blocks so the data keeps coming when other tools stop.

Built for AI pipelines

A Reddit scraper for LLM and AI pipelines

If you are pulling Reddit to feed a model, the hard part is not fetching, it is getting clean, consistent input. UGC Scraper returns one predictable JSON schema that drops straight into a RAG index or a fine-tuning set, no reshaping per source. Point it at the subreddits your users actually discuss, and hand the output to Claude or any LLM to summarize patterns, cluster pain points, or surface what is working.

Using Claude Desktop? Connect it directly over MCP with a two-line config (see the code tabs above) and let Claude scrape Reddit for you, no glue code.

Who it is for

One clean data stream, four kinds of work

AI datasets & RAG

Token-optimized input for LLM pipelines

Feed Reddit into a model and the hard part is clean, consistent input, not fetching. One predictable JSON schema drops straight into a RAG index or a fine-tuning set, no reshaping per source. Hand the output to Claude or any LLM to cluster pain points and summarize what users actually say.

Growth & lead research

A programmatic GummySearch alternative

Visual tools are great for manual browsing; an API is what you automate against. Pull posts from the subreddits your audience lives in, search by keyword, and turn the threads into structured signals your own tooling can tag, score, and act on.

Academic & linguistics

Bulk comment extraction without the plumbing

Text research needs volume across many threads and time periods. Skip the Python-library-and-proxy setup: pull the full comment tree per post and export CSV tables of author, body, score, and timestamp for analysis.

Archiving & media

Media links alongside the post payload

Each post carries its image and thumbnail URLs next to the text, so archival workflows can isolate, download, and catalog media without a second scrape or a separate crawler.

Drop-in integration

Three lines to your first scrape

No SDK to install. Call the endpoint with whatever you already use.

curl -X POST https://api.ugcscraper.com/v1/scrape \
  -H "Authorization: Bearer rps_live_..." \
  -H "Content-Type: application/json" \
  -d '{"url":"https://reddit.com/r/.../comments/..."}'

How it works

How the Reddit scraper works, in 3 steps

Get your API key

POST a Reddit URL

Send a post to /v1/scrape, or a subreddit, user, or search to their endpoint, with your key. That is the entire integration, and it never changes.

Get clean JSON back

The post and its full comment thread in one consistent schema, cached and ready to use. No HTML parsing, no div soup.

Pricing

Success-only pricing. Never pay for a blocked request.

No compute units, no separate proxy bill. Failed and blocked pulls are free, and only successful posts count toward your quota, cache hits included. Your quota resets at the start of each month; unused volume does not roll over.

Free

$0/mo

For prototypes and side projects.

1,000 successful posts / mo

Automatic fallback on every request
Consistent JSON schema
JSON & CSV export
Sub-second cache
Community support

Start free

What makes UGC Scraper the best Reddit scraper?

PRAW is great for building a bot; scraping is what you reach for when you need clean, resilient extraction at scale. Here is an objective, feature-by-feature look at the options. Rival cells reflect their public docs and pricing as of mid-2026; open-source tools vary, so those are "depends."

Tool	No API keys	No proxies to manage	Auto-fallback / reliability	Consistent JSON schema	Success-only pricing	Real free tier	CSV + JSON	LLM / MCP ready
UGC Scraper						(1k/mo)		(MCP)
Apify (trudax)		(configure proxies)		partial	(compute units + rental)	partial ($5 credits)		(MCP)
Outscraper				partial	(pay per post)	(50 posts)
RedScraper
PRAW / open-source	(PRAW) / depends	(you manage)		(you build it)	free	(free)	depends	(DIY)
Bright Data			partial		partial (billed-for-failures reported)	partial

Free and open-source tools win on price, but you operate everything and rebuild the schema yourself. Apify and Bright Data scale but add proxies, compute-unit billing, and complexity. UGC Scraper is the one that stays reliable without any of that, and only bills for data you actually get.

Where the alternatives fall short

The official API (PRAW)

PRAW is a clean wrapper for building bots, but it binds you to Reddit's OAuth app registration and its rate and pagination limits. A listing caps at about 1,000 items, and the terms and pricing around the official API have changed more than once. For deep extraction or an integration you do not want to babysit, it is the wrong tool.

Open-source scripts (YARS, Universal Reddit Scraper)

Free and flexible, but you own the whole operation. They hit public endpoints directly, so running them at any volume means renting and rotating your own residential proxies, handling blocks, and re-fixing the parser every time a layout shifts. The code is free; the infrastructure and maintenance are not.

Cloud actors (Apify, Outscraper)

These scale, but their billing is unpredictable: a fixed actor rental plus fluctuating compute and proxy metrics, and you can still be charged when an extraction fails midway. UGC Scraper bills only for successful posts, so a blocked pull costs you nothing.

FAQ

Reddit scraper FAQ

Will it break when Reddit changes its API or layout?

No. Every request runs through automatic fallback. When Reddit reprices the API, kills an endpoint, or blocks a path, the request reroutes and keeps working, so your integration never changes.

Do I need a Reddit API key to use UGC Scraper?

No. You sign up, get a UGC Scraper API key, and POST a Reddit URL. There is no Reddit OAuth app, developer approval, or credentials to manage.

Can it scrape Reddit comments, not just posts?

Yes. Every scrape returns the post plus its full comment thread with authors, scores, and timestamps, in one consistent JSON schema.

How do you get past Reddit's 1000-post listing limit?

Reddit listings cap at about 1000 items. UGC Scraper pulls the full comment tree per post, which is not bound by that cap, and keeps durable snapshots you can re-pull, so you get more than a single listing exposes.

What data do I get back, and in what formats?

The post and its full comment thread in one consistent JSON schema: title, author, score, comment count, and comments. Same shape no matter how it was fetched. Your stored history also exports as CSV (one row per snapshot, or one row per comment) for spreadsheets.

Will my scraper get rate-limited or IP-banned?

You never manage proxies or IPs. The fallback layer absorbs rate limits and blocks, so your own IP is never burned.

How are limits counted?

Only successful posts count toward your quota. Cache hits are included, and failed or blocked pulls are free. Your quota resets at the start of each month and unused volume does not roll over, so every month starts fresh.

Can I use it to feed Reddit data into an LLM or Claude?

Yes. The consistent JSON drops straight into RAG and fine-tuning pipelines and works well as a data source for LLM and agent workflows.

Is scraping Reddit legal?

UGC Scraper fetches publicly available Reddit content and requires no Reddit credentials. You are responsible for your own use and compliance, so consult your legal team for your specific use case.

Is there a free Reddit scraper tier?

Yes. The free plan includes 1,000 successful posts per month, the consistent schema, and the cache. Paid plans scale from there.

Can I use it as a programmatic GummySearch alternative?

Yes. GummySearch is a visual tool for manual audience research; UGC Scraper is the API version. Pull posts from the subreddits your users discuss or search by keyword, and pipe the structured output into your own monitoring, tagging, or analysis, at whatever scale you want.

Does the scraper break when Reddit changes its page layout?

A single-selector scraper breaks the moment a class name shifts. UGC Scraper uses automatic fallback, so when one path breaks another takes over and your integration never changes.

Can I use it with Claude Desktop over MCP?

Yes. Install the ugc-scraper-reddit-mcp server and add a two-line block to your Claude Desktop config with your API key. Claude can then scrape Reddit posts and comments and read your history directly, no glue code.

Reddit data that's still there tomorrow

One endpoint, one schema, one predictable price. The pipeline you set up once and stop worrying about. Start free; scale when your product does.

Get your API key View pricing