EveryInfra
Search⌘K
English
Sign inStart for free
EveryInfra

Everything AI needs to run.

Data scraping API, AI search, and CAPTCHA solving, with 413 data capabilities across 90 platforms.

Status

NewFree AI API for accounts that have topped up: organize the data you collect
Scroll down

What we offer

Collect content, comments, and products from supported platforms, perform AI searches, solve website CAPTCHAs.

Data crawling

Collect content, comments, and product data from supported platforms. Each item comes with fields and sources.

  • 90platforms
  • 413data capabilities
XiaohongshuDouyinTikTokInstagramYouTubeRedditAmazonGoogle Maps ReviewsLinkedInGlassdoorAirbnbBooking.comTripadvisorTrustpilotEtsyeBayBilibiliZhihuFacebookXThreadsSpotifyTaobaoShopee
Explore EveryData

Let AI organize the data you collect, free

Accounts that have topped up call the AI API at no charge: clean, deduplicate, classify, extract fields or summarize, in whatever format you need. It is OpenAI-compatible, so changing the base_url is enough.

  1. 50 requests/minTopped up, under $100 in totalUp to 5 at once
  2. 500 requests/min$100 or more in totalUp to 50 at once
  3. 5,000 requests/min$500 or more in totalUp to 500 at once

Calls are free and do not use your balance. Tiers follow your total top-ups; trial credit and credit codes do not count.

See the AI API Top up

The AI API is for organizing the data you collect with EveryInfra. If it is used for something else, or the calls look clearly abnormal, we may pause the AI API for that account; the account itself, its other APIs and its balance are not affected.

Call it with the OpenAI SDK. Default model gemini-3.6-flash; the full list is at /api/v1/models
from openai import OpenAI
import os

client = OpenAI(base_url="https://api.everyinfra.com/api/v1",
                api_key=os.environ["EVERYINFRA_API_KEY"])
rows = open("reviews.jsonl").read()  # data you collected with EveryInfra
prompt = "Deduplicate these reviews and label each one positive, negative or neutral, as JSON:"
reply = client.chat.completions.create(
    model="gemini-3.6-flash",
    messages=[{"role": "user", "content": prompt + "\n" + rows}],
)
print(reply.choices[0].message.content)

Capability examples

collection.json41 fields · Records 5
app_store12instagram9amazon8youtube6tiktok6
sourcerecordfieldtypevalue
app_store1557175442app_idnumber1557175442
instagramnasausernamestringnasa
amazonB0CFPHV9ZNtitlestringKindle Paperwhite 16GB
youtube@NASAnamestringNASA
tiktok@nasausernamestringnasa
app_store1557175442namestringObsidian - Connected Notes
instagramnasanamestringNASA
amazonB0CFPHV9ZNbrandstringAmazon
youtube@NASAhandlestring@NASA
tiktok@nasanamestringNASA
app_store1557175442ratingnumber4.5
app_store1557175442developerstringDynalist Inc.
app_store1557175442versionstring1.13.7
app_store1557175442rating_countnumber2722
app_store1557175442categorystringProductivity
app_store1557175442pricenumber0
app_store1557175442content_ratingstring4+
app_store1557175442size_displaystring39.6 MB
app_store1557175442min_osstringiOS 15.6
app_store1557175442urlstringhttps://apps.apple.com/us/app/obsidian-connected-notes/id1557175442
instagramnasafollowers_displaystring104M
instagramnasafollowing_countnumber91
instagramnasais_verifiedbooleantrue
instagramnasaposts_countnumber4931
instagramnasabiostringMaking the seemingly impossible, possible. ✨
instagramnasawebsitestringhttps://www.nasa.gov/
instagramnasaurlstringhttps://www.instagram.com/nasa/
amazonB0CFPHV9ZNcolorstringJade
amazonB0CFPHV9ZNasinstringB0CFPHV9ZN
amazonB0CFPHV9ZNstoragestring16 GB
amazonB0CFPHV9ZNdisplaystring7-inch
amazonB0CFPHV9ZNpricenull—
amazonB0CFPHV9ZNurlstringhttps://www.amazon.com/clp/B0CFPHV9ZN
youtube@NASAsubscribers_displaystring15.1M
youtube@NASAvideos_displaystring6.1K
youtube@NASAwebsitestringhttps://www.nasa.gov/
youtube@NASAurlstringhttps://www.youtube.com/@NASA/streams
tiktok@nasafollowers_displaystring1.8M
tiktok@nasalikes_displaystring9.4M
tiktok@nasabiostringMaking the seemingly impossible, possible.✨
tiktok@nasaurlstringhttps://www.tiktok.com/@nasa

How to integrate

Use EveryInfra in your existing AI tools or code.

Open the console

The fastest way in: give this prompt to your AI

Paste it into your AI assistant (Claude, ChatGPT and Cursor all work). It asks what you want to do, checks our catalogs, walks you through the plan and the cost, and only then writes code.

I want to use the EveryInfra API to collect platform data, search the web or solve CAPTCHAs. Act as my integration assistant and work in three steps.

Step 1: Don't write any code yet. Ask me questions over several rounds, one to three at a time, until all of this is clear:
1. What I want to do and what the results are for;
2. Which platforms or websites, whether the targets are keywords, links or accounts, and how many;
3. Which fields, how many rows, and the time range;
4. How to store the results, and where;
5. Which language the script uses, where it runs, and whether it runs once or on a schedule;
6. The most I want to spend on this.
If an answer is vague or contradicts an earlier one, ask again. If I say I don't know, suggest an option and ask me to confirm.

Step 2: Read EveryInfra's guide for AI agents at https://everyinfra.com/agents.md, then find the capabilities, parameters and prices in the catalogs (no key needed):
- Platform data: GET https://api.everyinfra.com/api/v1/social/catalog?platform=<name>
- Web search: GET https://api.everyinfra.com/api/v1/search/tools
- CAPTCHA: GET https://api.everyinfra.com/api/v1/captcha/types
Then walk me through the plan for approval: which capabilities and requests you will use, roughly how many calls, the estimated cost at catalog prices, and what the output will look like.

Step 3: Only after I approve, write the code and follow these API rules:
- Requests: platform data POST https://api.everyinfra.com/api/v1/social with the body {"platform", "action", "params"}; web search POST https://api.everyinfra.com/api/v1/search with {"tool", "params"}; CAPTCHA POST https://api.everyinfra.com/api/v1/captcha with {"type", ...other parameters}. Use parameter names and values exactly as the catalog lists them; do not guess.
- Auth: header Authorization: Bearer <API key>. Read the key from the EVERYINFRA_API_KEY environment variable; never put it in code, logs, or commits. If I don't have a key yet, remind me to create one at https://everyinfra.com/app#keys.
- Try one target with a small limit first and check the response shape before scaling up; a call with several targets is billed per target.
- For slow or large platform-data jobs, add "mode": "async" to the body. The API returns 202 and a job_id; poll GET https://api.everyinfra.com/api/v1/jobs/{job_id} until the status is succeeded or failed. A 202 is not a success. If a sync call comes back with a timeout error, resend the same request with "mode": "async".
- When a platform-data result carries next_page_token, pass it back as page_token for the next page with the other parameters unchanged; each page is one call.
- Failed or empty results are not charged in the end; any reserved amount goes back to the balance. Trust the billing receipt in the response.
- On a 422, fix the request using the candidate parameter names and allowed values in the error message; do not guess.
- On a 429 or 503, wait for the Retry-After delay and try again; if there is none, back off gradually.
- If the network drops or the outcome is unknown, keep the request_id / job_id and check it first. Do not resend automatically.

If you can connect to remote MCP servers, you can also connect to https://api.everyinfra.com/mcp (header Authorization: Bearer <API key>) and use its tools for the same work, with the same billing and rules as the API.

Every API page also has its own prompt, for example TikTok search

Select a tool to see the connection steps

Development

AI clients and plugins

Partners

Documentation, guides, and updates

Build · Engineering notes

View all build notes ↗
Build · Engineering notes
View all build notes ↗

Use what you need. See what it costs.

Each capability has its own billing unit. Check the price before making a call.

View pricing Browse the marketplace
EveryDataPer requestBilled by platform and collection method; a single request is not charged repeatedly for pagination
EverySolvePer successful resultBilled only when valid credentials are acquired; failed attempts incur no final charge
EverySearchPer requestBilled by search depth and tool tier; failed requests are not charged
AI APIFree after a top-upSee the AI API
Public list priceUSD
Estimated total$5.56

1,000 requests · $5.56 per 1,000

View Xiaohongshu

Estimated at public list price; nothing is charged. Your account's actual price is the one in the console.

Price and capability comparison

Prices and supported types are from each service's official pricing page on 2026-09-26; prices are per 1,000 billing units.

ServiceBilling unitEntry priceLowest public priceWhere we are ahead
EveryInfra (us)Pay per request, no monthly fee, failed requests are free
per 1,000 requests$0.6944$0.6944
SerperPrepaid, from $50.00
per 1,000 requests$1.00$0.30
  • 30% below its entry price
  • Lower minimum top-up
  • 17 search tools, including page reading, site crawling and cross-checking
  • One key for data, search and CAPTCHA solving
SerpApiMonthly plan from $25.00/monthFull comparison
per 1,000 requests$25.00$1.96
  • 64% below even its lowest price
  • No subscription; pay per request
  • One key for data, search and CAPTCHA solving
TavilyMonthly plan from $30.00/month
per 1,000 requests$7.50$5.00
  • 86% below even its lowest price
  • No subscription; pay per request
  • One key for data, search and CAPTCHA solving
ExaPay as you go
per 1,000 requests$7.00$7.00
  • 90% below even its lowest price
  • 17 search tools, including page reading, site crawling and cross-checking
  • One key for data, search and CAPTCHA solving
Brave Search APIPay as you go
per 1,000 requests$5.00$5.00
  • 86% below even its lowest price
  • 17 search tools, including page reading, site crawling and cross-checking
  • One key for data, search and CAPTCHA solving
  • EveryInfra (us)Pay per request, no monthly fee, failed requests are free
    Billing unit
    per 1,000 requests
    Entry price
    $0.6944
    Lowest public price
    $0.6944
  • SerperPrepaid, from $50.00
    Billing unit
    per 1,000 requests
    Entry price
    $1.00
    Lowest public price
    $0.30
    • 30% below its entry price
    • Lower minimum top-up
    • 17 search tools, including page reading, site crawling and cross-checking
    • One key for data, search and CAPTCHA solving
  • SerpApiMonthly plan from $25.00/monthFull comparison
    Billing unit
    per 1,000 requests
    Entry price
    $25.00
    Lowest public price
    $1.96
    • 64% below even its lowest price
    • No subscription; pay per request
    • One key for data, search and CAPTCHA solving
  • TavilyMonthly plan from $30.00/month
    Billing unit
    per 1,000 requests
    Entry price
    $7.50
    Lowest public price
    $5.00
    • 86% below even its lowest price
    • No subscription; pay per request
    • One key for data, search and CAPTCHA solving
  • ExaPay as you go
    Billing unit
    per 1,000 requests
    Entry price
    $7.00
    Lowest public price
    $7.00
    • 90% below even its lowest price
    • 17 search tools, including page reading, site crawling and cross-checking
    • One key for data, search and CAPTCHA solving
  • Brave Search APIPay as you go
    Billing unit
    per 1,000 requests
    Entry price
    $5.00
    Lowest public price
    $5.00
    • 86% below even its lowest price
    • 17 search tools, including page reading, site crawling and cross-checking
    • One key for data, search and CAPTCHA solving

Competitor prices are as listed on each service's official pricing page on 2026-09-26; later changes are not tracked. Data collection prices per 1,000 results assume only 10 results per call; most endpoints return up to 50–200 per call. Our prices are public list prices.

Frequently asked questions

What is EveryInfra?

EveryInfra is an infrastructure platform for AI applications. One API key gives you data collection (413 capabilities across 90 platforms), AI search, CAPTCHA solving, and an OpenAI-compatible AI API that is free after a top-up.

Browse the marketplace

Which platforms can I collect data from?

90 platforms, including Xiaohongshu (RedNote), Douyin, WeChat Official Accounts, TikTok, Amazon, Taobao, LinkedIn, and Google Maps reviews, across content and social, commerce, reviews, jobs and business, and web search. Each platform has its own API page with parameters and response fields.

See every platform by category

How is it billed, and am I charged for failures?

Each capability has its own price. Data collection costs $0.56–$5.56 per 1,000 requests, billed per request (a few list actions are priced per block of rows, noted next to their price), and invalid, failed, or empty requests are not charged. CAPTCHA solving costs $0.14–$4.58 per 1,000 successful solutions and is charged only on success.

View full pricing

Do I need to build scrapers or manage accounts and proxies?

No. You send a platform, a capability, and parameters, and get structured JSON back, without maintaining scrapers, platform accounts, or proxies yourself.

Can I use it from Claude, Cursor, or other AI tools?

Yes. The remote MCP server at https://api.everyinfra.com/mcp uses your API key, so clients such as Claude, Cursor, and ChatGPT can call every capability directly. Without MCP, you can call the REST API.

Set up MCP

Start building your next application

Start with a single capability and integrate infrastructure into your project.

Start for free

Explore the catalog without an account. Sign up to create an API key, then get a credit code on WeChat and redeem it in the console for trial credit, valid for 24 hours after redemption. No card is required to create an account.

Payment methods
  • VisaVisaMastercardAmerican ExpressCard
  • Apple Pay
  • Google Pay
  • Alipay
  • WeChat Pay
  • USDT (crypto)
  • PayNow
See payment steps