351 lines
12 KiB

__ __ __
/ // / ____ ____ / /_
/ // /_/ __ `/ _ \/ __/
/__ __/ /_/ / __/ /_
/_/ \__, /\___/\__/
+ Welcome to the 4get API documentation +
+ Terms of use
Do NOT misuse the API. Misuses can include... ::
1. Serp SEO scanning
2. Intensive scraping
3. Any other activity that isn't triggered by a human
4. Illegal activities in Canada
5. Constant "test" queries while developping your program
(please cache the API responses!)
Examples of good uses of the API ::
1. A chatroom bot that presents users with search results
2. Personal use
3. Any other activity that is initiated by a human
If you wish to engage in the activities listed under "misuses", feel
free to download the source code of the project and running 4get
under your own terms. Please respect the terms of use listed here so
that this website may be available to all in the far future.
P.s fuck whoever botted my site for months on end, choke on my dick
Get your instance running here ::
+ Passes
Depending of the instance, you may need to provide a "pass" token
in the cookies of your request. These can be obtained from solving
a captcha which will allow you to make 100 requests in the next 24
hours. In the future, you will be able to ask the serber maintainer
for a "pass" which will allow you to bypass the captcha requirement.
The captcha doesn't need javascript to work.
+ Decode the data
All payloads returned by the API are encoded in the JSON format. If
you don't know how to tackle the problem, maybe programming is not
for you.
All of the endpoints use the GET method.
+ Check if an API call was successful
All API responses come with an array index named "status". If the
status is something else than the string "ok", something went wrong.
You can supply the content of the "status" string back to your
application to inform the user of what went wrong.
The HTTP code will be 429 if your pass is invalid. It is set to 200
+ Get the next page of results
All API responses come with an array index named "npt". To get the
next page of results, you must make another API call with &npt.
Example ::
+ First API call
+ Second API call
/api/v1/web?npt=ddg1._rJ2hWmYSjpI2hsXWmYajJx < ... >
You shouldn't specify the search term, only the &npt parameter
The first part of the token before the dot (ddg1) refers to an
array position on the serber's memory. The second part is an
encryption key used to decode the data at that position. This way,
it is impossible to supply invalid pagination data and it is
impossible for a 4get operator to peek at the private data of the
user after a request has been made.
The tokens will expire as soon as they are used or after a 15
minutes inactivity period, whichever comes first.
+ Beware of null values!
Most fields in the API responses can return "null". You don't need
to worry about unset values.
+ API Parameters
To construct a valid request, you can use the 4get web interface
to craft a valid request, and replace "/web" with "/api/v1/web".
+ "date" and "time" parameters
"date" always refer to a calendar date.
"time" always refer to the duration of some media.
They are both integers that uses seconds as its unit. The "date"
parameter specifies the number of seconds that passed since January
1st 1970.
______ __ _ __
/ ____/___ ____/ /___ ____ (_)___ / /______
/ __/ / __ \/ __ / __ \/ __ \/ / __ \/ __/ ___/
/ /___/ / / / /_/ / /_/ / /_/ / / / / / /_(__ )
/_____/_/ /_/\__,_/ .___/\____/_/_/ /_/\__/____/
+ /ami4get
Tells you basic information about the 4get instance. CORS requests
are allowed on this endpoint.
+ /api/v1/web
+ &extendedsearch
When using the ddg(DuckDuckGo) scraper, you may make use of the
&extendedsearch parameter. If you need rich answer data from
additional sources like StackOverflow, music lyrics sites, etc.,
you need to specify the value of (string)"true".
The default value is "false" for API calls.
+ Parse the "spelling"
The array index named "spelling" contains 3 indexes ::
type: "including"
using: "4chan"
correction: '"4cha"'
The "type" may be any of these 3 values. When rendering the
autocorrect text inside your application, it should look like
what follows right after the parameter value ::
no_correction <Empty>
including Including results for %using%. Did you mean
not_many Not many results for %using%. Did you mean
As of right now, the "spelling" is only available on
+ Parse the "answer"
The array index named "answer" may contain a list of multiple
answers. The array index "description" contains a linear list of
nodes that can help you construct rich formatted data inside of
your application. The structure is similar to the one below:
title: "Higurashi"
type: "text"
value: "Higurashi is a great show!"
type: "quote"
value: "Source: my ass"
Each "description" node contains an array index named "type".
Here is a list of them:
+ title
+ quote
+ code
+ image
+ audio
Each individual node prepended with a "+" should be prepended by
a newline when constructing the rendered description object.
There are some nodes that differ from the type-value format.
Please parse them accordingly ::
+ link
type: "link"
url: "https://lolcat.ca"
value: "Visit my website!"
+ image
type: "image"
url: "https://lolcat.ca/static/pixels.png"
+ audio
type: "audio"
url: "https://lolcat.ca/static/whatever.mp3"
The array index named "table" is an associative array. You can
loop over the data using this PHP code, for example ::
foreach($table as $website_name => $url){ // ...
The rest of the JSON is pretty self explanatory.
+ /api/v1/images
All images are contained within "image". The structure looks like
below ::
title: "My awesome Higurashi image"
url: "https://lolcat.ca/static/profile_pix.png"
width: 400
height: 400
url: "https://lolcat.ca/static/pixels.png"
width: 640
height: 640
url: "https://tse1.mm.bing.net/th?id=OIP.VBM3BQg
width: 194
height: 160
The last image of the "source" array is always the thumbnail, and is
a good fallback to use when other sources fail to load. There can be
more than 1 source; this is especially true when using the Yandex
scraper, but beware of captcha rate limits.
+ /api/v1/videos
The "time" parameter for videos may be set to "_LIVE". For live
streams, the amount of people currently watching is passed in
+ /api/v1/news
Just make a request to "/api/v1/news?s=elon+musk". The payload
has nothing special about it and is very self explanatory, just like
the endpoint above.
+ /api/v1/music
Each entry under "song" contains a array index called "stream" that
looks like this ::
endpoint: sc
url: https://api-v2.soundcloud <...>
When the endpoint is something else than "linear", you MUST use
the specified endpoint. Otherwise, you are free to handle that
json+m3u8 crap yourself. If the endpoint is equal to "linear", the
URL should return a valid HTTP audio stream. To access the endpoint,
you must add the following prefix in your request, like so:
+ /favicon
Get the favicon for a website. The only parameter is "s", and must
include the protocol for fetching in case the favicon is not cached
Example ::
If we had to revert to using Google's favicon cache, it will throw
an error in the X-Error header field. If Google's favicon cache
also failed to return an image, or if you're too retarded to specify
a valid domain name, a default placeholder image will be returned
alongside the "404" HTTP error code.
+ /proxy
Get a proxied image. Useful if you don't want to leak your user's IP
address. The parameters are "i" for the image link and "s" for the
Acceptable "s" parameters:
portrait 90x160
landscape 160x90
square 90x90
thumb 236x180
cover 207x270
original <Original resolution>
You can also ommit the "s" parameter if you wish to view the
original image. When an error occurs, an "X-Error" header field
is set.
+ /audio/linear
Get a proxied audio file. Does not support "Range" headers, as it's
only used to proxy small files (hence why it's called linear DUH)
The parameter is "s" for the audio link.
+ /audio/sc
Get a proxied audio file for SoundCloud. Does not support downloads
trough WGET or CURL, since it returns 30kb~160kb "206 Partial
Content" parts, due to technical limitations that comes with
converting m3u8 playlists to seekable audio files. If you use this
endpoint, you must support these 206 codes and also handle the
initial 302 HTTP redirect. I used this method as I didn't want to
store information about your request needlessly. This method also
allows noJS users to access the files.
The parameter is "s" for the SoundCloud JSON m3u8 abomination. It
does not support "normal" SoundCloud URLs at this time.
+ /audio/spotify
Get a proxied Spotify audio file. Accepts a track ID for the "s"
parameter. Will only allow you to fetch the 30 second preview since
I don't feel like fucking with cookies and accounts every fucking
living moment of my life. You must handle the initial 302 redirect
to the /audio/linear endpoint.
+ Appendix
If you have any questions or need clarifications, please send an
email my way to will at lolcat.ca