LurkAPI

Live public posts, profiles, ads, comments and transcripts from 16 social platforms and the web.

호스팅형 MCP 서버

npx add-mcp 'https://api.lurkapi.com/mcp'

Claude Code, Codex, Cursor 등에 설치됩니다

문서

Quickstart

LurkAPI fetches public posts, profiles, ads and comments from Facebook, Reddit, YouTube and 13 more platforms, as clean data. Use it from an AI app like Claude, from your own code, or right here in the browser.

Web page: https://lurkapi.com/docs

Ways to start:

Try it now (no signup): Run a real request on this page and see what comes back. 5 free tries a day. Run a request
Ask your AI app (no code): Add LurkAPI to Claude, ChatGPT, Cursor or VS Code. Then ask for what you want in plain English. Add to Claude
Call it from code (for developers): Send one request, get JSON back. You start with 100 free credits, no card needed. Get a free API key

Try it live

Pick what to fetch and press Run request. It's real data, and you don't need an account: 5 free tries a day.

Live playground: https://lurkapi.com/docs#try

Use it in your AI app

Add LurkAPI to your AI app once. Then ask in plain English, like “Which ads is Allbirds running right now?” Your AI picks the right request, runs it and answers with what it found.

It works through MCP, the standard way AI apps connect to outside tools. Any app that supports MCP can use it, not just these.

Claude

Opens Claude with LurkAPI filled in. Click Add, then Connect to sign in. In a chat, switch LurkAPI on from the tools menu.

Add to Claude

Add it by hand:

In Claude, open Settings → Connectors and choose Add custom connector.
Name it LurkAPI, paste https://api.lurkapi.com/mcp and click Add.
Click Connect and sign in to LurkAPI, or create an account in the same step.
In a chat, switch LurkAPI on from the tools menu.

Free Claude plans allow one custom connector; on Team plans an owner adds it.

ChatGPT

Open Plugins, click + and choose Add custom MCP server. Paste this address, pick OAuth and sign in.

https://api.lurkapi.com/mcp?compact

Cursor

Cursor asks to install LurkAPI, then you sign in. No key to paste.

Add to Cursor

VS Code

Adds LurkAPI to Copilot Chat. You sign in the first time it starts.

Add to VS Code

Claude Code

Run this in your terminal to install the LurkAPI plugin: the server plus skills for whole research jobs (below). Ask away before you sign in: Claude Code gets 5 free tries a day on 1-credit endpoints. To sign in, type /mcp in Claude Code, pick plugin:lurkapi:lurkapi and choose Authenticate.

claude plugin marketplace add mesmerlord/lurkapi-claude-plugin && claude plugin install lurkapi@lurkapi

Only the server, or for scripts and CI:

Add the server without the skills:

Terminal:

claude mcp add --transport http lurkapi https://api.lurkapi.com/mcp

For scripts and CI, skip the sign-in and send a key in a header instead:

Terminal:

claude mcp add --transport http lurkapi https://api.lurkapi.com/mcp \
  --header "x-api-key: YOUR_API_KEY"

Any other AI app

Paste this address. An app that signs in with OAuth needs nothing else.

https://api.lurkapi.com/mcp

If your app can't sign in:

Give it your key in a header. Most apps take a JSON config like this:

mcp.json:

{
  "mcpServers": {
    "lurkapi": {
      "url": "https://api.lurkapi.com/mcp",
      "headers": {
        "x-api-key": "YOUR_API_KEY"
      }
    }
  }
}

If it only takes an address, add your key to the end: https://api.lurkapi.com/mcp?key=YOUR_API_KEY.

Each app signs in to your LurkAPI account and gets its own key. Its calls spend your credits, the same as calls from code. To disconnect an app, revoke its key on your dashboard.

Skills

Skills turn a whole job into one ask: lurk a brand, tear down its ads, research Reddit, vet a creator, read a channel's videos or size up demand. Each one knows the calls to chain and what they cost. The Claude Code plugin includes them; Codex, Cursor and other agents that read Agent Skills can add them on their own, and call the API with your key in LURKAPI_API_KEY. See the skills.

Terminal:

npx skills add mesmerlord/lurkapi-claude-plugin

Advanced options

Fewer tools. Your app sees one tool per endpoint, named after it, like facebook_search_ads. Each endpoint's page shows its tool name and a prompt to try. To keep the list short, name the platforms you want: https://api.lurkapi.com/mcp?platforms=facebook,reddit. A platform's id is the end of its docs address, like facebook in /docs/facebook.
Two tools instead of many. Some apps, like ChatGPT, load every tool up front and cap how many they take. For them, https://api.lurkapi.com/mcp?compact gives two tools: find_endpoints to search the catalog and call_endpoint to run one. Prices are the same.
Trimmed results. To save tokens, tool results leave out empty fields, keep one URL per image or video, and give transcripts as plain text unless you pass timestamps=true. A result over 45,000 characters comes in parts. Its _note gives the offset for the next part, and later parts are free for 15 minutes. Pass trim=false for the full response.
Recipes. Prompts for common jobs: ad_research, brand_snapshot, vet_creator, subreddit_research and channel_transcripts. In Claude Code they're slash commands, like /lurkapi:ad_research allbirds.com; other apps that support MCP prompts list them in their prompt menu.
Spend caps. Tools whose price varies (web pages, speech-to-text) take max_credits: a call that would cost more stops before it spends anything.
Errors your AI can fix. A failed call says whether retrying can help (retryable), what to change (fix) and, for a misnamed tool or parameter, the corrected call (did_you_mean).

Your first API call

Three steps. You need a terminal, or a few lines of code in Node.js, Python, Go, PHP or Ruby.

1. Get a free API key

An API key is a password for your code. You send it with every request, so we know the call is yours.

Get a free API key

2. Send a request

A request is a web address that says what you want, sent with your key. This one runs Search Ad Library ads with an example search. Replace YOUR_API_KEY with your key, then paste the curl version into a terminal:

curl:

curl "https://api.lurkapi.com/v1/facebook/adLibrary/search/ads?query=running+shoes&country=US&status=active" \
  -H "x-api-key: YOUR_API_KEY"

JavaScript:

const params = new URLSearchParams({
  query: "running shoes",
  country: "US",
  status: "active",
});
const res = await fetch(`https://api.lurkapi.com/v1/facebook/adLibrary/search/ads?${params}`, {
  headers: { "x-api-key": process.env.LURKAPI_API_KEY },
});
const data = await res.json();
if (!data.success) throw new Error(`${data.code}: ${data.error}`);
console.log(data.searchResults);

Python:

import os
import requests

res = requests.get(
    "https://api.lurkapi.com/v1/facebook/adLibrary/search/ads",
    params={
        "query": "running shoes",
        "country": "US",
        "status": "active",
    },
    headers={"x-api-key": os.environ["LURKAPI_API_KEY"]},
    timeout=60,
)
data = res.json()
if not data["success"]:
    raise RuntimeError(f"{data['code']}: {data['error']}")
print(data["searchResults"])

The code versions read your key from LURKAPI_API_KEY, an environment variable, so the key stays out of your code. Pick a language once and every page shows it.

3. Read the result

You get JSON back: plain text with named fields, which any language can read. Every answer starts the same way:

success is true when the call worked.
credits_remaining is your balance after it.
credits_charged is what this call cost.
Then comes the data, here under searchResults.

200 OK:

{
  "success": true,
  "credits_remaining": 4985,
  "credits_charged": 1,
  "searchResults": […]
}

Each endpoint's page lists its parameters and every field it returns, with a real example response and a live playground. They're all listed under All endpoints.

Authentication

Send your key in the x-api-key header. Authorization: Bearer <key> works too.

Keys start with lk_live_. Each one is shown once, when you create it. Create and revoke keys on your dashboard.

AI apps that sign in through LurkAPI get their own key, so there's nothing to paste. An app that can't sign in can send a key in a header, or add it to the address as ?key=.

Keep your key secret. Call LurkAPI from a server or a script, never from a web page your visitors can inspect.

Agents: getting a key needs a human sign-in. Ask your user to sign in at https://lurkapi.com/login and paste the key from the dashboard.

Requests and responses

Every endpoint takes a GET request. Parameters go in the query string, the part of the address after ?.

Unknown parameters are ignored. Blank ones count as unset. If you repeat a parameter, the first value wins.

Answers are JSON. A successful one has success: true, credits_charged, credits_remaining and the data, as shown above. A failed one has success: false and a code to branch on (see Errors).

Credits

Each call costs credits. Endpoints cost 1–31 credits per call, and each endpoint's page shows its price.

Without an account: 5 free tries a day from the playground, the free tools, or Claude Code before you sign in.
New accounts: 100 free credits, no card needed.
Every day: if your balance is under 25, it's topped back up to 25.
Credit packs: one-off purchases, no subscription. Credits never expire.
Auto top-up: opt in on your dashboard and a pack is bought with your saved card whenever your balance drops below a threshold you pick (at most 3 a day and 10 a month), so calls don't stop on a 402. Every purchase has an invoice there.

Pack	Credits	Price	Per 1,000 credits
Starter	10,000	$9.00	$0.90
Growth	100,000	$59.00	$0.59
Scale	500,000	$225.00	$0.45

What you pay for

Charged: every call that reached the platform. That includes not_found and other errors from the endpoint itself, like a bad cursor.
Free: calls we stop before that, like a missing or invalid parameter, a bad key (401), too few credits (402) or too many requests (429). Failures on our side (5xx) are free too.
Cached answers cost the same as fresh ones. Free tries without an account follow the same rules.
Cap a call with max_credits. Any call takes it, and one that would cost more stops with 402 over_max_credits and costs nothing. It matters where the price varies: web pages and speech-to-text transcripts.

If your balance is too low for a call, you get 402 insufficient_credits and nothing is charged. Credit balance is free, so you can check your balance before a big batch.

Errors

When a call fails, you get an HTTP status of 400 or higher and a body like this:

400 Bad Request:

{
  "success": false,
  "error": "query: required",
  "code": "invalid_params",
  "docs": "https://lurkapi.com/docs/facebook/search-ads",
  "issues": [
    {
      "path": "query",
      "message": "required"
    }
  ]
}

Branch on code. The error text is written for people and may change. docs links to the page that covers it, and parameter errors list each problem in issues.

Over MCP, errors carry three more fields for your AI. A failed call says whether retrying can help (retryable), what to change (fix) and, for a misnamed tool or parameter, the corrected call (did_you_mean).

Retry 429, 502 and 503. A 429 or 503 comes with a retry-after header: the number of seconds to wait.

Errors caught before we reach the platform are free, and so are failures on our side. A not_found is charged. See what you pay for.

Status	Code	Meaning
400	`invalid_params`	A parameter is missing or invalid. `issues` names each one.
401	`missing_api_key`	No API key was sent. Put it in the `x-api-key` header.
401	`invalid_api_key`	This key doesn't exist or was revoked.
401	`invalid_token`	Your playground session expired. Reload the page.
402	`insufficient_credits`	Not enough credits for this call. Buy a pack, or wait for tomorrow's free top-up.
402	`over_max_credits`	This call would cost more than the `max_credits` you set. The error says what it needs; raise `max_credits` to allow it. Not charged.
403	`account_suspended`	This account is suspended. Email support@lurkapi.com. Not charged.
403	`not_public`	The platform shows this account or post only to people who are logged in. Try a public one. Charged, since the platform was checked.
404	`not_found`	What you asked for (a post, an ad, a subreddit) doesn't exist, or there's no endpoint at that path. Charged if the platform was checked.
404	`transcript_unavailable`	The video has no captions, and speech-to-text heard no speech in it. You pay the base credit, never the speech-to-text extra. The answer is cached.
405	`method_not_allowed`	Endpoints only take `GET`, with parameters in the URL.
429	`rate_limited`	Too many calls in a short time. Wait as many seconds as the `retry-after` header says, then retry.
429	`signup_required`	Today's free tries are used up. Sign up for free daily credits.
500	`internal_error`	Something broke on our side. Retry. Not charged.
502	`upstream_error`	The platform didn't send a usable answer. Retry. Not charged.
503	`upstream_busy`	All our connections to the platform are busy. Retry in a few seconds. Not charged.
504	`upstream_timeout`	The platform took over 30 seconds to answer (90 for web pages). Retry. Not charged.

Rate limits

Each account can make 50 requests every 10 seconds. That's 5 a second, sustained.

Go over it and you get 429 rate_limited with a retry-after header. Wait that many seconds, then retry.

High-volume accounts get 100 requests every 10 seconds. Email support@lurkapi.com to raise your limit.

Without an account, the playground and free tools allow 5 requests per IP address a day. After that you get 429 signup_required.

Pagination

List endpoints return one page of results per call. For the next page, send the same request again with the cursor from the answer. It's called cursor or after, and each endpoint's page says which.

On the last page the cursor is null. Each page costs the endpoint's credits.

Caching and freshness

We keep recent answers for a while, so asking again comes back fast. How long depends on how quickly the data changes, from 30 seconds to 30 days. Each endpoint's page shows its own time.

A cached answer costs the same credits as a fresh one.

Endpoint	Cached for up to
Facebook: Search Ad Library ads	30 minutes
Facebook: Ad Library ads by company	30 minutes
Facebook: Ad Library ad details	6 hours
Facebook: Search Ad Library companies	1 day
Facebook: Public profile	10 minutes
Facebook: Public post	5 minutes
Facebook: Profile posts	10 minutes
Facebook: Profile events	1 hour
Facebook: Post comments	5 minutes
Reddit: Subreddit posts	1 minute
Reddit: Subreddit details	1 hour
Reddit: Subreddit search	1 minute
Reddit: Search Reddit	1 minute
Reddit: Post comments	1 minute
YouTube: Video details	1 hour
YouTube: Video transcript	30 days
YouTube: Video comments	10 minutes
YouTube: Comment replies	10 minutes
YouTube: Channel details	6 hours
YouTube: Channel videos	1 hour
YouTube: Channel Shorts	1 hour
YouTube: Search YouTube	15 minutes
TikTok: Profile	15 minutes
TikTok: Video details	15 minutes
TikTok: Video transcript	30 days
TikTok: Profile videos	5 minutes
TikTok: Video comments	5 minutes
TikTok: Comment replies	5 minutes
TikTok: Hashtag videos	10 minutes
TikTok: Song	1 hour
TikTok: Song videos	10 minutes
TikTok: Search TikTok Ad Library	1 hour
TikTok: TikTok Ad Library ad	3 hours
TikTok: Live status	30 seconds
TikTok: Public stories	1 minute
TikTok: TikTok Shop product	5 minutes
TikTok: Search videos	10 minutes
TikTok: TikTok Shop store products	15 minutes
TikTok: TikTok Shop search suggestions	1 day
Google: Search Google advertisers	1 day
Google: Google ads by company	12 hours
Google: Google ad details	12 hours
Google: Search Google Maps	12 hours
Google: Google Maps place details	1 day
Google: Google Maps reviews	1 hour
Google: Search Google News	10 minutes
Google: Google autocomplete	6 hours
Bluesky: Profile	30 seconds
Bluesky: User posts	3 minutes
Bluesky: Post	3 minutes
Snapchat: Profile and stories	15 minutes
Truth Social: Profile	10 minutes
Truth Social: User posts	2 minutes
Truth Social: Post	5 minutes
Telegram: Channel	1 hour
Telegram: Channel posts	5 minutes
Telegram: Post	10 minutes
X (Twitter): Tweet	5 minutes
X (Twitter): Profile	1 hour
X (Twitter): User tweets	5 minutes
Threads: Threads profile	1 hour
Threads: Threads account posts	1 minute
Threads: Threads post	5 minutes
Instagram: Profile	10 minutes
Instagram: User posts	3 minutes
Instagram: Post	5 minutes
Instagram: Post comments	2 minutes
Instagram: User reels	3 minutes
Instagram: Media transcript	30 days
Instagram: Search reels	10 minutes
Instagram: Similar profiles	1 hour
LinkedIn: LinkedIn profile	6 hours
LinkedIn: LinkedIn company page	1 hour
LinkedIn: LinkedIn company posts	1 hour
LinkedIn: Search LinkedIn jobs	1 hour
LinkedIn: LinkedIn job details	1 hour
LinkedIn: Search LinkedIn ads	1 hour
LinkedIn: LinkedIn ad details	1 hour
Link-in-bio: Linktree page	1 day
eBay: Search eBay listings	15 minutes
eBay: eBay listing details	15 minutes
Web: Scrape a page (Auto)	not cached
Web: Fetch a page (plain request)	not cached
Web: Render a page (browser)	not cached
Web: Unblock a page (separate IPs)	not cached
Web: Fetch a page from a residential IP	not cached
Web: Render a page from a residential IP	not cached
Web: Scrape a page with your own steps	not cached

All endpoints

Every endpoint, with its path, its price and its tool name in AI apps.

Facebook

Endpoint	Path	Cost	MCP tool
Search Ad Library ads	`/v1/facebook/adLibrary/search/ads`	1 credit	`facebook_search_ads`
Ad Library ads by company	`/v1/facebook/adLibrary/company/ads`	1 credit	`facebook_company_ads`
Ad Library ad details	`/v1/facebook/adLibrary/ad`	1 credit	`facebook_ad`
Search Ad Library companies	`/v1/facebook/adLibrary/search/companies`	1 credit	`facebook_search_companies`
Public profile	`/v1/facebook/profile`	1 credit	`facebook_profile`
Public post	`/v1/facebook/post`	1 credit	`facebook_post`
Profile posts	`/v1/facebook/profile/posts`	1 credit	`facebook_profile_posts`
Profile events	`/v1/facebook/profile/events`	1 credit	`facebook_profile_events`
Post comments	`/v1/facebook/post/comments`	1 credit	`facebook_post_comments`

Endpoint	Path	Cost	MCP tool
Subreddit posts	`/v1/reddit/subreddit`	1 credit	`reddit_subreddit_posts`
Subreddit details	`/v1/reddit/subreddit/details`	1 credit	`reddit_subreddit_details`
Subreddit search	`/v1/reddit/subreddit/search`	1 credit	`reddit_subreddit_search`
Search Reddit	`/v1/reddit/search`	1 credit	`reddit_search`
Post comments	`/v1/reddit/post/comments`	1 credit	`reddit_post_comments`

YouTube

Endpoint	Path	Cost	MCP tool
Video details	`/v1/youtube/video`	1 credit	`youtube_video`
Video transcript	`/v1/youtube/video/transcript`	1 credit	`youtube_transcript`
Video comments	`/v1/youtube/video/comments`	1 credit	`youtube_comments`
Comment replies	`/v1/youtube/video/comment/replies`	1 credit	`youtube_comment_replies`
Channel details	`/v1/youtube/channel`	1 credit	`youtube_channel`
Channel videos	`/v1/youtube/channel-videos`	1 credit	`youtube_channel_videos`
Channel Shorts	`/v1/youtube/channel/shorts`	1 credit	`youtube_channel_shorts`
Search YouTube	`/v1/youtube/search`	1 credit	`youtube_search`

TikTok

Endpoint	Path	Cost	MCP tool
Profile	`/v1/tiktok/profile`	1 credit	`tiktok_profile`
Video details	`/v2/tiktok/video`	1 credit	`tiktok_video`
Video transcript	`/v1/tiktok/video/transcript`	1 credit; 2 if speech-to-text runs	`tiktok_transcript`
Profile videos	`/v3/tiktok/profile/videos`	1 credit	`tiktok_profile_videos`
Video comments	`/v1/tiktok/video/comments`	1 credit	`tiktok_comments`
Comment replies	`/v1/tiktok/video/comment/replies`	1 credit	`tiktok_comment_replies`
Hashtag videos	`/v1/tiktok/search/hashtag`	1 credit	`tiktok_hashtag_videos`
Song	`/v1/tiktok/song`	1 credit	`tiktok_song`
Song videos	`/v1/tiktok/song/videos`	1 credit	`tiktok_song_videos`
Search TikTok Ad Library	`/v1/tiktok/ad-library/search`	1 credit	`tiktok_ad_library_search`
TikTok Ad Library ad	`/v1/tiktok/ad-library/ad`	1 credit	`tiktok_ad_library_ad`
Live status	`/v1/tiktok/user/live`	1 credit	`tiktok_live`
Public stories	`/v1/tiktok/user/stories`	1 credit	`tiktok_stories`
TikTok Shop product	`/v1/tiktok/product`	1 credit	`tiktok_product`
Search videos	`/v1/tiktok/search/keyword`	1 credit	`tiktok_search_keyword`
TikTok Shop store products	`/v1/tiktok/shop/products`	1 credit	`tiktok_shop_products`
TikTok Shop search suggestions	`/v1/tiktok/shop/search/suggestions`	1 credit	`tiktok_shop_suggestions`

Google

Endpoint	Path	Cost	MCP tool
Search Google advertisers	`/v1/google/adLibrary/advertisers/search`	1 credit	`google_search_advertisers`
Google ads by company	`/v1/google/company/ads`	1 credit	`google_company_ads`
Google ad details	`/v1/google/ad`	1 credit	`google_ad`
Search Google Maps	`/v1/google/maps/search`	1 credit	`google_maps_search`
Google Maps place details	`/v1/google/maps/place`	1 credit	`google_maps_place`
Google Maps reviews	`/v1/google/maps/reviews`	1 credit	`google_maps_reviews`
Search Google News	`/v1/google/news`	1 credit	`google_news`
Google autocomplete	`/v1/google/autocomplete`	1 credit	`google_autocomplete`

Bluesky

Endpoint	Path	Cost	MCP tool
Profile	`/v1/bluesky/profile`	1 credit	`bluesky_profile`
User posts	`/v1/bluesky/user/posts`	1 credit	`bluesky_user_posts`
Post	`/v1/bluesky/post`	1 credit	`bluesky_post`

Snapchat

Endpoint	Path	Cost	MCP tool
Profile and stories	`/v1/snapchat/profile`	1 credit	`snapchat_profile`

Truth Social

Endpoint	Path	Cost	MCP tool
Profile	`/v1/truthsocial/profile`	1 credit	`truthsocial_profile`
User posts	`/v1/truthsocial/user/posts`	1 credit	`truthsocial_user_posts`
Post	`/v1/truthsocial/post`	1 credit	`truthsocial_post`

Endpoint	Path	Cost	MCP tool
Channel	`/v1/telegram/channel`	1 credit	`telegram_channel`
Channel posts	`/v1/telegram/channel/posts`	1 credit	`telegram_channel_posts`
Post	`/v1/telegram/post`	1 credit	`telegram_post`

X (Twitter)

Endpoint	Path	Cost	MCP tool
Tweet	`/v1/twitter/tweet`	1 credit	`twitter_tweet`
Profile	`/v1/twitter/profile`	1 credit	`twitter_profile`
User tweets	`/v1/twitter/user-tweets`	1 credit	`twitter_user_tweets`

Threads

Endpoint	Path	Cost	MCP tool
Threads profile	`/v1/threads/profile`	1 credit	`threads_profile`
Threads account posts	`/v1/threads/user/posts`	1 credit	`threads_user_posts`
Threads post	`/v1/threads/post`	1 credit	`threads_post`

Instagram

Endpoint	Path	Cost	MCP tool
Profile	`/v1/instagram/profile`	1 credit	`instagram_profile`
User posts	`/v2/instagram/user/posts`	1 credit	`instagram_user_posts`
Post	`/v1/instagram/post`	1 credit	`instagram_post`
Post comments	`/v2/instagram/post/comments`	1 credit	`instagram_post_comments`
User reels	`/v1/instagram/user/reels`	1 credit	`instagram_user_reels`
Media transcript	`/v2/instagram/media/transcript`	1 credit; 2 if speech-to-text runs	`instagram_media_transcript`
Search reels	`/v2/instagram/reels/search`	1 credit	`instagram_reels_search`
Similar profiles	`/v1/instagram/profile/similar`	1 credit	`instagram_similar_profiles`

Endpoint	Path	Cost	MCP tool
LinkedIn profile	`/v1/linkedin/profile`	1 credit	`linkedin_profile`
LinkedIn company page	`/v1/linkedin/company`	1 credit	`linkedin_company`
LinkedIn company posts	`/v1/linkedin/company/posts`	1 credit	`linkedin_company_posts`
Search LinkedIn jobs	`/v1/linkedin/jobs/search`	1 credit	`linkedin_search_jobs`
LinkedIn job details	`/v1/linkedin/job`	1 credit	`linkedin_job`
Search LinkedIn ads	`/v1/linkedin/ads/search`	1 credit	`linkedin_search_ads`
LinkedIn ad details	`/v1/linkedin/ad`	1 credit	`linkedin_ad`

Link-in-bio

Endpoint	Path	Cost	MCP tool
Linktree page	`/v1/linktree`	1 credit	`linkinbio_linktree`

eBay

Endpoint	Path	Cost	MCP tool
Search eBay listings	`/v1/ebay/search`	2 credits	`ebay_search`
eBay listing details	`/v1/ebay/item`	2 credits	`ebay_item`

Web

Endpoint	Path	Cost	MCP tool
Scrape a page (Auto)	`/v1/web/scrape`	1 credit; +3 if a browser on separate IPs gets it, +4 if it needs a residential IP, +13 if a browser renders it on a residential IP, +10 if a captcha is solved	`web_scrape`
Fetch a page (plain request)	`/v1/web/fetch`	1 credit	`web_fetch`
Render a page (browser)	`/v1/web/render`	1 credit; 11 if a captcha is solved	`web_render`
Unblock a page (separate IPs)	`/v1/web/unblock`	4 credits	`web_unblock`
Fetch a page from a residential IP	`/v1/web/residential`	5 credits	`web_residential`
Render a page from a residential IP	`/v1/web/render-residential`	14 credits; 24 if a captcha is solved	`web_render_residential`
Scrape a page with your own steps	`/v1/web/custom`	1 credit; +3 if a browser on separate IPs gets it, +4 if it needs a residential IP, +13 if a browser renders it on a residential IP, +10 if a captcha is solved	`web_custom`

Account

Endpoint	Path	Cost	MCP tool
Credit balance	`/v1/account/credit-balance`	free	`account_credit_balance`

For AI agents

Base URL: https://api.lurkapi.com
Auth: x-api-key: lk_live_… header (or Authorization: Bearer lk_live_…). Get a free key: 100 free credits on signup, no card.
Requests: GET with query parameters. Responses are JSON with success, credits_remaining and the data at the top level; errors add code and docs.
Price: Endpoints cost 1–31 credits per call, from $0.45 per 1,000 credits. Calls turned away before they reach the platform (invalid parameters, a key or balance problem, a rate limit) are free, and so are 5xx errors. A not_found is charged, since the platform was checked. How charging works. Every response has credits_charged and credits_remaining; GET /v1/account/credit-balance is free.
Rate limit: 50 requests per 10 seconds per account (100 for high-volume accounts: support@lurkapi.com); over it, 429 rate_limited with retry-after.
Agents: getting a key needs a human sign-in. Ask your user to sign in at https://lurkapi.com/login and paste the key from the dashboard.
MCP server: https://api.lurkapi.com/mcp (Streamable HTTP). Every endpoint is a tool. Add platforms to keep the list short: https://api.lurkapi.com/mcp?platforms=facebook,reddit gives one tool per endpoint of those platforms, plus the free account tool. Platform ids: facebook, reddit, youtube, tiktok, google, bluesky, snapchat, truthsocial, telegram, twitter, threads, instagram, linkedin, linkinbio, ebay and web. For a client that loads every tool up front and caps how many it takes, like ChatGPT, https://api.lurkapi.com/mcp?compact gives two tools instead: find_endpoints to search the catalog and call_endpoint to run one, charged the same. To save tokens, tool results leave out empty fields, keep one URL per image or video, and give transcripts as plain text unless you pass timestamps=true. A result over 45,000 characters comes in parts. Its _note gives the offset for the next part, and later parts are free for 15 minutes. Pass trim=false for the full response. Tools whose price varies (web pages, speech-to-text) take max_credits: a call that would cost more stops before it spends anything. A failed call says whether retrying can help (retryable), what to change (fix) and, for a misnamed tool or parameter, the corrected call (did_you_mean). Prompts for common jobs: ad_research, brand_snapshot, vet_creator, subreddit_research and channel_transcripts. In Claude Code they're slash commands, like /lurkapi:ad_research allbirds.com; other apps that support MCP prompts list them in their prompt menu.
Machine-readable: /openapi.json, /llms.txt, /llms-full.txt, and every docs page as markdown at its URL + .md.

GET https://api.lurkapi.com/ returns a JSON index of every endpoint.

Support

Email support@lurkapi.com about anything: a wrong result, a broken endpoint, billing or your account. Tell us the endpoint and the parameters you sent. Never send your key.

The status page shows whether each endpoint is working right now.

LurkAPI

호스팅형 MCP 서버

문서

Quickstart

Try it live

Use it in your AI app

Claude

ChatGPT

Cursor

VS Code

Claude Code

Any other AI app

Skills

Advanced options

Your first API call

1. Get a free API key

2. Send a request

3. Read the result

Authentication

Requests and responses

Credits

What you pay for

Errors

Rate limits

Pagination

Caching and freshness

All endpoints

Facebook

Reddit

YouTube

TikTok

Google

Bluesky

Snapchat

Truth Social

Telegram

X (Twitter)

Threads

Instagram

LinkedIn

Link-in-bio

eBay

Web

Account

For AI agents

Support