/imageboard

An easy uniform wrapper over the popular imageboards' API

Primary LanguageJavaScriptMIT LicenseMIT

imageboard

Provides easy programmatic access to the popular imageboards.

Supported imageboard engines:

  • makaba — 2ch.hk's proprietary engine.

    1. 2ch.hk — demo

  • 4chan — 4chan.org's proprietary engine.

    1. 4chan.org — demo

  • vichan — An open-source 4chan-compatible engine running on PHP/MySQL whose development started in 2012. The codebase has seen various maintainers take over and then leave off over the years, but as of late 2023, it seems like it's still being maintained and receiving new features.

    1. tvch.moe — demo
    2. tahta.ch — demo
    3. vecchiochan.com — demo
    4. diochan.com
    5. soyjak.party
    6. vichan.pl
    7. wizchan.org

  • lainchan — a fork of vichan that adds a bit of features. Is no longer in development since 2023.

    1. lainchan.org — demo
    2. leftypol.org

  • OpenIB (formerly infinity) — A 2013 fork of vichan engine with the goal of supporting an "infinite" amount of user-managed boards as opposed to a finite set of predefined boards. No longer maintained since 2018.

    1. 8kun.top (formerly 8ch.net) — demo
    2. smuglo.li — demo

  • lynxchan — An alternative engine Node.js/MongoDB whose development started in 2015. Rather than mimicking any existing engine, it set off on its own path and ended up becoming a popular choice (of its time) provided that there's really not much else to choose from. Some choices made by the author are questionable and the overall approach doesn't look professional to me. For example, the engine has a bunch of quite obvious but easily-fixable issues that the author refuses to recognize and has no interest in fixing. The author's demeanor, in general, is somewhat controversial and not to everyone's liking.

    1. kohlchan.net — demo
    2. endchan.net — demo
    3. alogs.space — demo
    4. bandada — demo

  • jschan — An alternative engine written in Node.js/MongoDB whose development started in 2019. Isn't really adopted by anyone, perhaps because there haven't been any new imageboards since its development has started. Compared to lynxchan, purely from a technical perspective, it looks much more professional and mature, and the author is a well-known developer.

    1. junkuchan.org — demo
    2. jakparty.soy — demo
    3. heolkek.cafe
    4. 27chan.org
    5. niuchan.org
    6. zzzchan.xyz
    7. trashchan.xyz
    8. 94chan.org
    9. ptchan.org

Originally created as part of the anychan imageboard GUI.

Details on each imageboard engine API

Thoughts

Features

  • (optional) Parses comments HTML markup into Content JSON structures.
  • (optional) Automatically generates shortened "previews" for long comments.
  • (optional) Automatically inserts quotes in replying comments.
  • (optional) Automatically builds a tree of replies: each comment has a list of "in-reply-to" comments and a list of "replying" comments.
  • (optional) Automatically generates thread title when it's missing.
  • (not on all imageboards) Provides the API to:
    • Create threads
    • Post comments
    • Report comments
    • Request and solve a CAPTCHA
    • Log In / Log Out

Install

npm install imageboard --save

Example

Here's an example for Node.js, although this library also works in a web browser.

P.S. The full source code for this example can be found in test/test.js file.

Create a new folder and cd into it.

This example will be using 4chan.org as a data source.

Create a file called fourChan.js and paste the following code into it.

import Imageboard from 'imageboard/node'

export default Imageboard('4chan')

Next, create an index.js file and paste the following code into it. The code prints the first ten of 4chan.org boards:

import fourChan from './fourChan.js'

// Prints the first 10 boards.
fourChan.getBoards().then(({ boards }) => {
  const boardsList = boards.slice(0, 10).map(({
    id,
    title,
    category,
    description
  }) => {
    return `* [${category}] /${id}/ — ${title} — ${description}`
  })
  console.log(boardsList.join('\n'))
})

The output:

* [Creative] /3/ — 3DCG — "/3/ - 3DCG" is 4chan's board for 3D modeling and imagery.
* [Japanese Culture] /a/ — Anime & Manga — "/a/ - Anime & Manga" is 4chan's imageboard dedicated to the discussion of Japanese animation and manga.
* [Other] /adv/ — Advice — "/adv/ - Advice" is 4chan's board for giving and receiving advice.
... the other boards ...

Now, add the code that prints the first five threads on 4chan.org /a/ board:

import { getCommentText } from 'imageboard'

// Prints the first five threads on `/a/` board.
fourChan.getThreads({
  boardId: 'a'
}).then(({ threads, board }) => {
  const threadsList = threads.slice(0, 5).map(({
    id,
    title,
    createdAt,
    commentsCount,
    attachmentsCount,
    comments
  }) => {
    return [
      `#${id}`,
      createdAt,
      title,
      getCommentText(comments[0]) || '(empty)',
      `${commentsCount} comments, ${attachmentsCount} attachments`
    ].join('\n\n')
  })
  console.log(threadsList.join('\n\n~~~~~~~~~~~~~~~~~~~~~~~~~\n\n'))
})

The output:

#193605320

Wed Sep 25 2019 20:48:54 GMT+0300 (Moscow Standard Time)

Facebook announces Horizon, a VR massive-multiplayer world

DATABASE DATABASE JUST LIVING IN THE DATABE WO-OH

12 comments, 5 attachments

~~~~~~~~~~~~~~~~~~~~~~~~~

... the other threads ...

Now, add the code that prints the first five comments of a certain thread:

import { getCommentText } from 'imageboard'

// Prints the first five comments of thread #193605320 on `/a/` board.
fourChan.getThread({
  boardId: 'a',
  threadId: 193605320
}).then(({ thread, board }) => {
  const commentsList = thread.comments.slice(0, 5).map((comment) => {
    const {
      id,
      title,
      createdAt,
      replies,
      attachments
    } = comment
    const parts = []
    parts.push(`#${id}`)
    parts.push(createdAt)
    if (title) {
      parts.push(title)
    }
    parts.push(getCommentText(comment) || '(empty)')
    if (attachments) {
      parts.push(`${attachments.length} attachments`)
    }
    if (title && replies) {
      parts.push(`${replies.length} replies`)
    }
    return parts.join('\n\n')
  })
  console.log(commentsList.join('\n\n~~~~~~~~~~~~~~~~~~~~~~~~~\n\n'))
})

The output:

#193605320

Wed Sep 25 2019 20:48:54 GMT+0300 (Moscow Standard Time)

Facebook announces Horizon, a VR massive-multiplayer world

DATABASE DATABASE JUST LIVING IN THE DATABE WO-OH

1 attachments

3 replies

~~~~~~~~~~~~~~~~~~~~~~~~~

#193605743

Wed Sep 25 2019 21:02:00 GMT+0300 (Moscow Standard Time)

«DATABASE DATABASE JUST LIVING IN THE DATABE WO-OH»
Fuck you OP I saw horizon and now I am mad that it’s bait

~~~~~~~~~~~~~~~~~~~~~~~~~~

... the other comments ...

API

Imageboard(idOrConfig, options)

The default exported function. Creates a new imageboard instance.

There're three different exports:

  • imageboard — The default (envirnoment-agnostic) one. Requires specifying a custom sendHttpRequest parameter.
  • imageboard/browser — The one that can be used in a web browser. Has a web-browser-specific sendHttpRequest function "built in".
  • imageboard/node — The one that can be used in Node.js. Has a Node.js-specific sendHttpRequest function "built in".
import Imageboard from 'imageboard/node'

const fourChan = Imageboard('4chan', options)

If the imageboard is supported by this library out-of-the-box, such imageboard's id can be passed as the first argument:

  • "2ch"
  • "27chan"
  • "4chan"
  • "8ch" (8kun.top)
  • "94chan"
  • "alogs.space"
  • "arisuchan"
  • "bandada"
  • "diochan"
  • "endchan"
  • "heolcafe"
  • "jakparty.soy"
  • "junkuchan"
  • "kohlchan"
  • "lainchan"
  • "leftypol"
  • "niuchan"
  • "ptchan"
  • "smugloli"
  • "tahtach"
  • "tvchan"
  • "vecchiochan"
  • "wizardchan"
  • "zzzchan"

Otherwise, i.e. when the imageboard is not supported by the library out-of-the-box, the first argument should be an imageboard config object.

Sidenote: Making HTTP requests in a web browser will most likely require a CORS proxy.

None of the imageboards (4chan.org, 8kun.top, 2ch.hk, etc) allow calling their API from other websites in a web browser: they're all configured to block Cross-Origin Resource Sharing (CORS), so a CORS proxy is required in order for a third party website to be able to query their API.

To bypass CORS limitations, the request() function would have to call fetch() not with the original imageboard URL like https://imageboard.net/boards.json directly but rather with a "proxied" URL like https://my-cors-proxy-address.net/?url=${encodeURIComponent('https://imageboard.net/boards.json')}.

How to set up a CORS proxy.

Available options:

  • commentUrl?: string — A template to use when formatting the url property of type: "post-link"s (links to other comments) in parsed comments' content. Is "/{boardId}/{threadId}#{commentId}" by default.

  • threadUrl?: string — A template to use when formatting the url property of type: "post-link"s (links to threads) in parsed comments' content. Is "/{boardId}/{threadId}" by default.

  • messages?: Messages — "Messages" ("strings", "labels") used when parsing comments content. See Messages.

  • commentLengthLimit?: number — A number telling the maximum comment length (in "points" which can be thought of as "characters and character equivalents for non-text content") upon exceeding which a preview is generated for a comment (as comment.contentPreview).

  • useRelativeUrls?: boolean — Determines whether to use relative or absolute URLs for attachments. Relative URLs are for the cases when an imageboard is temporarily hosted on an alternative domain and so all attachments are too meaning that the default imageboard domain name shouldn't be present in attachment URLs. Is false by default.

See a list of "advanced" options

  • parseContent?: boolean — Is true by default. Can be set to false to skip parsing comment HTML into Content. The rationale is that when there're 500-some comments in a thread parsing all of them up-front can take up to a second on a modern desktop CPU which results in subpar user experience. By deferring parsing comments' HTML an application could first only parse the first N comments' HTML and only as the user starts scrolling would it proceed to parsing the next comments. Or maybe a developer wants to use their own HTML parser or even render comments' HTML as is. If parseContent is set to false then each non-empty comment will have their content being the original unmodified HTML string. In such cases thread.title won't be autogenerated when it's missing.

  • addParseContent?: boolean — Pass true to add a .parseContent() method and a .createContentPreview() method to each comment. This could be used in scenarios when parseContent: false option is passed, for example, to only parse comments as they become visible on screen as the user scrolls rather than parsing all the comments in a thread up-front. When passing addParseContent: true option, also pass expandReplies: true option, otherwise .parseContent() won't go up the chain of quoted comments.

    • The .parseContent() function receives an optional object containing the parameters for the function:

      • getCommentById(id): comment? — Returns a comment by an id. This parameter should only be passed if some of the comment objects of a thread might have been replaced by the application (for whatever reason) since the time of receiving the original thread object.

        • Why might an application choose to replace some of the thread's comments? An example would be refreshing the thread by first fetching a new thread2 object and then updating the list of the comments in the original thread with the new comments from the newly fetched thread2. But in that case, some of the older comments' replies count (.replies.length) might've changed, so those older comments' .replies lists should be updated, and then those older comments should be re-rendered in order to reflect the updated replies count. And re-rendering a comment when using React framework would mean simply replacing those comment objects with copies of themselves, which is called updating their "object reference".

    • When addParseContent: true option is passed, each comment will also have a .hasContentBeenParsed() function that can be used to find out whether the comment's content has been parsed.

    • When addParseContent: true option is passed, each comment will also have an .onContentChange() function that should be called whenever the comment's perceived .content changes for any reason. An example would be marking a comment as "hidden": in that case, the application might decide to update of that comment's replies by replacing their quotes of this comment with ">Hidden comment" placeholder text.

      • The .onContentChange() function receives an optional object containing the parameters for the function:

        • getCommentById(id): comment? — Returns a comment by an id. This parameter should only be passed if some of the comment objects of a thread might have been replaced by the application (for whatever reason) since the time of receiving the original thread object. See the notes on .parseContent()'s getCommentById parameter for an example on why might an application choose to replace some of the comment objects in some scenarios.

  • expandReplies?: boolean — Every comment has optional properties: comment.replyIds[] and comment.inReplyToIds[]. If expandReplies: true parameter is passed, it will add additional optional properties on every comment: comment.replies[] and comment.inReplyTo[] that're arrays of Comments rather than arrays of just Comment.ids. By default, expandReplies: false is used to prevent JSON circular structure: this way a whole thread could be serialized into a *.json file.

  • getPostLinkProperties?: (comment?: Comment) => object — Returns an object all properties of which will be added to the post-link object. The comment argument will be undefined if the comment was removed.

  • getPostLinkText?: (postLink: object) => string? — A function that can be used to define the text content of post-links. It should return either a text for a post-link, or undefined, in which case it's simply ignored. An example of how it might be used:

// The application might choose to display `post-link`s to all
// "hidden" comments as links with "Hidden comment" text inside.
getPostLinkText(postLink) {
  // If the `post` has not been removed.
  // If the application has marked the post as "hidden".
  if (postLink.meta.postIsHidden) {
    return "Hidden comment"
  }
}

// Returns additional `post-link` properties.
// These properties don't get refreshed in `comment.onContentChange()`.
getPostLinkProperties(comment) {
  return {
    // Whether the application has marked the `comment` object as "hidden".
    postIsHidden: comment.hidden
  }
}

  • getSetCookieHeaders?: ({ headers }) => string[] — Returns a list of Set-Cookie header values for an HTTP response. The default implementation just returns headers.getSetCookie(), but a developer could supply their own implementation, for example, to work around web browsers' limitations as they don't expose set-cookie response header.

  • When parsing comments, it automatically generates quotes when a comment references another comment in the same thread. The options for generating such quotes are:

    • generatedQuoteMaxLength?: number — The maximum length of an auto-generated quote. This is an approximate value.
    • generatedQuoteMinFitFactor: number? — Sets the minimum preferrable length of an auto-generated quote at generatedQuoteMinFitFactor * generatedQuoteMaxLength.
    • generatedQuoteMaxFitFactor: number? — Sets the maximum preferrable length of an auto-generated quote at generatedQuoteMaxFitFactor * generatedQuoteMaxLength.
    • generatedQuoteGetCharactersCountPenaltyForLineBreak?: ({ textBefore: string }) => number — Returns the character count equivalent for a "line break" (\n) character. The idea is to "tax" multi-line texts when trimming by max length. By default, having \n characters in text is not penalized in any way and those characters aren't counted.
    • minimizeGeneratedPostLinkBlockQuotes?: boolean — One can pass true to indicate that auto-generated quotes are minimized by default until the user expands them manually. This would mean that auto-generated quotes shouldn't be accounted for when calculating the total length of a comment when creating a shorter "preview" for it in case it exceeds the maxum preferred length.

imageboard methods

supportsFeature(feature: string)

Tells whether the imageboard supports a certain feature.

feature argument could be one of:

  • "getBoards"
  • "getTopBoards"
  • "findBoards"
  • "getThread.withLatestComments"
  • "getThreads.sortByRatingDesc"
  • "findThreads"
  • "findComments"
  • "findComments.boardId" — If findComments() supports searching for comments across all threads of a given board.
  • "findComments.threadId" — If findComments() supports searching for comments in a given thread.
  • "rateComment"
  • "reportComment"
  • "createThread"
  • "updateThread"
  • "createComment"
  • "updateComment"
  • "createBlockBypass"
  • "getCaptcha"
  • "logIn"
  • "logIn.tokenPassword"
  • "logOut"

Returns:

  • true if the feature is supported by the imageboard.
  • false otherwise.

getBoards()

Fetches a list of boards on an imageboard.

Returns an object with properties:

For some imageboards, the list could be huge. For example, 8ch.net (8kun.top) has about 20,000 boards, and to get just the "top 20 boards" one could use getTopBoards() function instead.

getTopBoards()

Fetches a list of "top" boards on an imageboard.

Returns an object with properties:

For example, 8ch.net (8kun.top) has about 20,000 boards, so getTopBoards() returns just the "top 20 boards" while getBoards() returns all 20,000 boards.

This function is only supported if supportsFeature('getTopBoards') returns true.

For example, 8ch.net (8kun.top) has about 20,000 boards, so supportsFeature('getTopBoards') returns true for that specific imageboard and getTopBoards() returns just the "top 20 boards".

findBoards()

Searches for a (non-full) list of boards matching a search query. For example, if an imageboard supports creating "user boards", and there're a lot of them, then getBoards() should return just the most popular ones, and to discover all other boards, searching by a query should be used.

Parameters object:

  • search: string — Search query. Could be an empty string.

Returns an object with properties:

This function isn't currently implemented in any of the supported imageboard engines. To see if a given imageboard supports this function, use supportsFeature('findBoards') function.

getThreads()

Fetches a list of threads on a board.

Parameters object:

  • boardId: string — Board ID.

  • withLatestComments?: boolean — Pass true to get latest comments for each thread. Latest comments are added as thread.latestComments[], but not necessarily for all threads in the list, because the result might differ depending on the imageboard engine. For example, 4chan provides latest comments for all threads in the list as part of its "get threads list" API response, while other imageboard engines don't — which is lame — and so the latest comments have to somehow be fetched separately by, for example, going through the "pages" of threads on a board. When doing so, some threads might not get their latestComments[] at all, resulting in thread.latestComments[] being undefined. That might happen for different reasons. One reason is that the list of threads on a board changes between the individual "read page" requests, so some threads might get lost in this space-time gap. Another reason is that it wouldn't be practical to go through all of the "pages" because that'd put unnecessary load on the server, and that's when withLatestCommentsMaxPages parameter comes into effect.

    • comments[] will be an array of a single element — the "root" comment of the thread.
    • latestComments[] will not include the "root" comment of the thread.

  • withLatestCommentsMaxPages?: number — The maximum number of threads list "pages" to fetch in order to get the latest comments for each thread. When the engine doesn't provide the latest comments for each thread as part of its generic "get threads list" API response, the latest comments are fetched by going through every "page" of threads on a board. Therefore, to get the latest comments for all threads on a board, the code has to fetch all "pages", of which there may be many (for example, 10). At the same time, usually imageboards warn the user of the "max one API request per second" rate limit (which, of course, isn't actually imposed). So the code decides to play it safe and defaults to fetching just the first "page" of threads to get just those threads' latest comments, and doesn't set the latest comments for the rest of the threads. A developer may specify a larger count of "pages" to be fetched. It would be safe to specify "over the top" (larger than actual) pages count because the code will just discard all "Not found" errors. All pages are fetched simultaneously for better UX due to shorter loading time, so consider web browser limits for the maximum number of concurrent HTTP requests when setting this parameter to a high number.

  • commentLengthLimitForWithLatestComments?: number — Same as commentLengthLimit but for thread.latestComments.

  • sortBy?: string — Could be used to sort the list of threads by something. For example, makaba engine supports sortBy: "rating-desc".

  • Any other properties here will override the corresponding properties of the options object that was passed to the Imageboard() function when creating an imageboard instance.

    • defaultAuthorName?: string — If board object is available and has a board.features.defaultAuthorName property, it could be passed here to parse the default author name as an empty one.

Returns an object with properties:

findThreads()

Searches for a (non-full) list of threads on a given board matching a search query.

Parameters object:

  • boardId: string — Board ID

  • search: string — Search query. Could be an empty string.

  • Any other properties here will override the corresponding properties of the options object that was passed to the Imageboard() function when creating an imageboard instance.

    • defaultAuthorName?: string — If board object is available and has a board.features.defaultAuthorName property, it could be passed here to parse the default author name as an empty one.

Returns an object with properties:

  • threads — a list of Threads
  • board — (optional) Board info

This function isn't currently implemented in any of the supported imageboard engines. To see if a given imageboard supports this function, use supportsFeature('findThreads') function.

getThread()

Fetches a thread.

Parameters object:

  • boardId: string — Board ID.

  • threadId: number — Thread ID.

  • archived?: boolean — (optional) Pass true when requesting an archived thread. This flag is not required in any way, but, for makaba engine, it reduces the number of HTTP Requests from 2 to 1 because in that case it doesn't have to attempt to read the thread by a non-"archived" URL (which returns 404 Not Found) before attempting to read it by an "archived" URL.

  • afterCommentId?: number — (optional) (experimental) Could be used to only fetch comments after a certain comment.

  • afterCommentNumber?: number — (optional) (experimental) Could be used to only fetch comments after a certain comments count (counting from the first comment in the thread).

  • Any other properties here will override the corresponding properties of the options object that was passed to the Imageboard() function when creating an imageboard instance.

    • defaultAuthorName?: string — If board object is available and has a board.features.defaultAuthorName property, it could be passed here to parse the default author name as an empty one.

Returns an object with properties:

findComments()

Searches for a (non-full) list of comments in a given thread matching a search query.

Parameters object:

  • boardId: string — Board ID.
  • threadId?: number — Thread ID. If not specified then searches across all threads on the board.
  • search: string — Search query. Could be an empty string.

Returns an object with properties:

  • comments — a list of Comments
  • thread — (optional) Thread info
  • board — (optional) Board info

This function isn't currently implemented in any of the supported imageboard engines. To see if a given imageboard supports this function, use supportsFeature('findComments') function.

rateComment()

Some imageboards (like 2ch.hk) allow upvoting or downvoting threads and comments on certain boards (like /po/litics on 2ch.hk).

Parameters object:

  • boardId: string — Board ID.
  • threadId: number — Thread ID.
  • commentId: number — Comment ID.
  • up: boolean — Upvote or downvote.

Returns:

  • true — The vote has been accepted.
  • false — The vote has not been accepted. For example, if the user has already voted.

createThread()

Creates a new thread on a board.

Parameters object:

  • boardId: string — Board ID.
  • accessToken?: string — If the user is authenticated, put an accessToken here to bypass CAPTCHA.
  • authorEmail?: string — Comment author email.
  • authorName?: string — Comment author name.
  • attachments?: File[] — Attachments.
  • title?: string — Comment title (subject).
  • content?: string — Comment content (text).
  • makaba-specific properties:
    • authorIsThreadAuthor?: boolean — "Comment author is the thread author" flag.
    • authorIconId?: number | string — Comment author icon ID.
    • tags?: string[] — Thread tags.
    • captchaType?: string — CAPTCHA type. Possible values: "2chcaptcha".
    • captchaId?: string — CAPTCHA ID.
    • captchaSolution?: string — CAPTCHA solution.

Returns an object:

  • id: number — Thread ID

updateThread()

Updates a thread.

Parameters object:

  • boardId: string — Board ID.
  • threadId: number — Thread ID.
  • Same other properties as in createThread()

This function isn't currently implemented in any of the supported imageboard engines. To see if a given imageboard supports this function, use supportsFeature('updateThread') function.

createComment()

Creates a new comment in a thread.

Parameters object:

  • boardId: string — Board ID.
  • threadId: number — Thread ID.
  • accessToken?: string — If the user is authenticated, put an accessToken here to bypass CAPTCHA.
  • authorEmail?: string — Comment author email.
  • authorName?: string — Comment author name.
  • attachments?: File[] — Attachments.
  • title?: string — Comment title (subject).
  • content?: string — Comment content (text).
  • makaba-specific properties:
    • authorIsThreadAuthor?: boolean — "Comment author is the thread author" flag.
    • authorIconId?: number | string — Comment author icon ID.
    • captchaType?: string — CAPTCHA type. Possible values: "2chcaptcha".
    • captchaId?: string — CAPTCHA ID.
    • captchaSolution?: string — CAPTCHA solution.

Returns an object:

  • id: number — Comment ID

updateComment()

Updates a comment.

Parameters object:

  • boardId: string — Board ID.
  • threadId: number — Thread ID.
  • commentId: number — Comment ID.
  • Same other properties as in createComment()

This function isn't currently implemented in any of the supported imageboard engines. To see if a given imageboard supports this function, use supportsFeature('updateComment') function.

getCaptcha()

Requests a CAPTCHA,

Parameters object:

  • boardId?: string — Board ID.
  • threadId?: number — Thread ID, if posting a comment in a thread.

Returns an object:

  • captcha: object:
    • id: string — CAPTCHA id, can be used to get CAPTCHA image URL.
    • type: string — CAPTCHA type. Possible values: "text".
    • challengeType: string — CAPTCHA type. Possible values: "image", "image-slider".
    • characterSet?: string — A name of a character set for a "text" CAPTCHA solution. For example, "numeric" or "russian".
    • expiresAt: Date — CAPTCHA expiration date.
    • image: object — CAPTCHA image:
      • type: string — image mime-type.
      • url: string — image URL. May be relative or absolute.
      • width: number — image width.
      • height: number — image height.
    • backgroundImage?: object — "image-slider" CAPTCHA background image:
      • type: string — image mime-type.
      • url: string — image URL. May be relative or absolute.
      • width: number — image width.
      • height: number — image height.
  • canRequestNewCaptchaAt?: Date — When the user is allowed to request a new CAPTCHA.

createBlockBypass()

Engines such as jschan or lynxchan introduce a concept of a "block bypass". A "block bypass" is a token that allows its owner to "bypass" a "block" that the server may have imposed on them. Such a "block" could block the user from posting comments or threads or submitting reports.

In order to prevent spam, users are untrusted by default, and are required to solve a CAPTCHA.

Offtopic: Although in the modern age of AI development automatically solving CAPTCHAs has become an easy task, so I'd personally advise utilizing other means of spam prevention. For example, kohlchan.net requires a user to provide a "Proof-of-Work" in a form of a result of a lengthy computation. And popular imageboards like 4chan.org or 2ch.hk provide their users an option to bypass CAPTCHA by purchasing a "pass" which could be considered a "Proof-of-Stake".

This createBlockBypass() method issues a new "block bypass" token.

Parameters object:

  • captchaId: string — Captcha ID.
  • captchaSolution: string — Captcha solution.

Returns an object:

  • token: string — "Block bypass" token.
  • expiresAt: date — The date when the "block bypass" token expires.

logIn()

Performs a login.

Parameters object:

  • token: string — Login token. For example, 4chan calls them "passes".
  • tokenPassword?: string — Login token password. For example, 4chan uses them (it calls it a "PIN").

Returns an object:

  • accessToken?: string — Access token.

logOut()

Performs a logout.

No parameters.

Returns undefined.

reportComment()

Reports a comment or a thread.

Parameters object:

  • boardId: string — Board ID.
  • threadId: number — Thread ID.
  • commentId: number — Comment ID.
  • accessToken?: string — If the user is authenticated, put an accessToken here to bypass CAPTCHA.
  • content?: string — Report content. Is used in makaba engine.
  • reasonId?: number | string — Report reason ID. Is used in 4chan engine.
  • legalViolationReasonId?: number — Report reason ID (31) in case of a violation of United States law. No reasonId should be passed. Is only used in 4chan engine.
  • captchaId?: string — CAPTCHA ID. Is used in 4chan engine.
  • captchaSolution?: string — CAPTCHA solution. Is used in 4chan engine.

Returns undefined.

Miscellaneous API

The following functions are exported for "advanced" use cases. In other words, they're being used in anychan and that's the reason why they're exported.

getConfig(id: string): object?

Returns an imageboard config by its id. Example: getConfig("4chan").

Can be used in cases when an application for whatever reasons needs to know the imageboard info defined in the *.json file, such as domain, engine, etc.

getCommentText(comment: Comment, options: object?): string?

Generates a textual representation of Comment's content. This is just a getPostText() function re-exported from social-components for convenience.

Is used in the examples in this document.

Available options (optional argument):

  • messages: Messages? — (optional) "Messages" ("strings", "labels") used when generating text from content. See Messages.
  • skipPostQuoteBlocks: boolean? — (optional) Set to true to skip all "block" (not "inline") "comment quotes" (quotes citing a particular comment).
  • skipGeneratedPostQuoteBlocks: boolean? — (optional) Set to true to skip all "block" (not "inline") autogenerated "comment quotes" (autogenerated quotes citing a particular comment).

sortThreadsWithPinnedOnTop(threads: Thread[]): Thread[]

Sorts a list of threads so that pinned ones are on top according to their pinnedOrder.

createHttpRequestFunction()

Creates a custom sendHttpRequest() function that could be passed when creating an Imageboard() using the default export from the imageboard package.

See HTTP request function of this document for more details.

supportsFeature()

Tells if the feature is supported by the imageboard. See the docs for the instance method supportsFeature(feature) of Imageboard.

Arguments:

  • imageboardIdOrConfig — Imageboard ID string or configuration object.
  • feature — Feature name.

Returns:

  • true if the feature is supported by the imageboard
  • false otherwise

Models

Board

See Board.md or ./types/index.d.ts.

Thread

See Thread.md or ./types/index.d.ts.

Comment

See Comment.md or ./types/index.d.ts.

Content

Each comment could have a content property, which is optional. The content could be simple text or it could be "rich"-formatted text in which case it uses some kind of a "markup" language like Markdown or HTML.

In order to convert that "markup" language into a Content structure, the comment's content has to be "parsed". This package does that by default. To turn off the automatic parsing feature, pass parseContent: false parameter.

When parsing a comment's content, if commentLengthLimit parameter was passed, it will automatically create a contentPreview if content surpasses the commentLengthLimit. One could also use a comment.createContentPreview({ maxLength }) method to create a preview of a content with a custom maxLength limit.

Attachment

An attachment can be a:

Embedded Link Attachments

This library doesn't parse links to YouTube/Twitter/etc inside comments. Instead, this type of functionality is offloaded to a separate library. For example, anychan uses loadResourceLinks() and expandStandaloneAttachmentLinks() from social-components library when rendering comments to load YouTube/Twitter/etc links and embed the attachments directly in comments.

Messages

Sometimes an optional messages object can be passed to define "messages" ("strings", "labels") used when parsing comments content.

Messages used for setting the content of comment links:

{
  ...,
  comment: {
    default: "Comment",
    deleted: "Deleted comment",
    external: "Comment from another thread"
  },
  thread: {
    default: "Thread"
  }
}

Messages used when generating content text (autogenerated quotes, autogenerated thread title):

{
  ...,
  textContent: {
    ... // See the "Messages" section in the readme of `social-components` package.
  }
}

See ./types/Messages.d.ts for the most up-to-date description of the Messages structure.

Adding a new imageboard

Follow the instructions in docs/add-new-imageboard.md.

Adding a new engine

Follow the instructions in docs/add-new-engine.md.

HTTP request function

When importing Imageboard function from imageboard package directly rather than from a sub-package like imageboard/browser or imageboard/node, one would have to supply a sendHttpRequest() parameter:

sendHttpRequest() — (required) Sends HTTP requests to an imageboard's API.

  • parameters argument:
    • method: string — HTTP request method, in upper case.
    • url: string — HTTP request URL.
    • body?: object — POST request body JSON object.
    • headers?: object — HTTP request headers, with header names in lower case.
    • cookies?: object — HTTP request cookies.
  • Returns a Promise:
    • When successful, the Promise should resolve to an object:
      • url: string — HTTP request URL. If there were any redirects in the process, this should be the final URL that it got redrected to.
      • status: number — HTTP response status.
      • responseText?: string — HTTP response text. Can be omitted if there's no response content.
      • headers: object — HTTP response headers. For example, the headers property of a fetch() response. Must be an object having functions:
        • .get(headerName: string) — Returns the value of an HTTP response header.
        • .getSetCookie() — Returns a list of set-cookie header values in HTTP response.
    • Otherwise, in case of an error, the Promise should reject with an Error instance having optional properties:
      • url: string — HTTP request URL. If there were any redirects in the process, this should be the final URL that it got redrected to.
      • status: number — HTTP response status.
      • responseText?: string — HTTP response text. Can be omitted if there's no response content.
      • headers: object — HTTP response headers. For example, the headers property of a fetch() response. Must be an object having functions:
        • .get(headerName: string) — Returns the value of an HTTP response header.
        • .getSetCookie() — Returns a list of set-cookie header values in HTTP response.
await sendHttpRequest({
  method: "GET",
  url: "https://8kun.top/boards.json"
}) === {
  url: "https://8kun.top/boards.json",
  status: 200,
  headers: {},
  responseText: "[
    { "uri": "b", "title": "Random" },
    ...
  ]"
}
Sidenote: Rationale for returning responseText rather than just response.

Because the response could be in different forms:

  • application/json
  • text/html
  • text/plain, when something weird happens like the server being down.

So it's not always JSON, and having a response parameter be sometimes of this type and sometimes of that type wouldn't be considered a professional approach.

Sidenote: Rationale for returning status, url and headers.

  • status and url are used when reading archived threads on 2ch.hk imageboard. When requesting an archived thread on 2ch.hk, it always redirects with a 302 status to a URL that looks like /boardId/arch/yyyy-mm-dd/res/threadId.json. As one can see, there's the archivedAt date in that final "redirect-to" URL, and it's nowhere else to be read. So this library has to have the access to the final URL after any redirects.

  • headers are used when parsing an imageboard API's response to a "log in" request: this library attempts to read the authentication cookie value from that response in order to return it to the application code.

An appropriate sendHttpRequest() function could be created using the exported createHttpRequestFunction() function:

import { createHttpRequestFunction } from 'imageboard'

const sendHttpRequest = createHttpRequestFunction({
  fetch,
  FormData
})

Required parameters of createHttpRequestFunction():

Optional parameters of createHttpRequestFunction():

  • setHeaders({ headers }) — Sets any custom HTTP request headers.
  • attachCookiesInWebBrowser({ cookies, headers }) — Attaches any custom cookies to HTTP request when running in a web browser environment.
  • getRequestUrl({ url }) — Transforms HTTP request URL in any arbitrary way.
  • getResponseStatus({ status, headers }) — Returns any custom HTTP response status.
  • getFinalUrlFromResponse({ url, headers }) — Returns any custom "final" URL for HTTP response.
  • mode — Overrides the mode parameter of the fetch() function.
  • credentials — Overrides the credentials parameter of the fetch() function.
  • redirect — Overrides the redirect parameter of the fetch() function. The default is "manual".

Imageboards API

4chan

vichan

vichan engine was originally a fork of Tinyboard engine having more features. In November 2017, vichan engine was put in "no longer being maintained" state, although later, around 2022, a group of volunteers seems to have picked up the development from the former maintainer.

Chans running on vichan:

  • lainchan.org — Runs on a custom fork of vichan called lainchan

infinity

infinity is a vichan fork permitting users to create their own boards (hence the name). As of April 2017, infinity engine is no longer being maintained. OpenIB is a security-focused fork of the infinity engine which is no longer being maintained too.

Chans running on infinity:

makaba

The only imageboard running on makaba engine is currently 2ch.hk.

lynxchan

lynxchan seems to be the only still-being-maintained imageboard engine left. Has JSON API and POST API.

Chans running on lynxchan:

jschan

jschan is an alternative engine written in Node.js/MongoDB whose development started in 2019. Isn't really adopted by anyone, perhaps because there haven't been any new imageboards since its development has started. Compared to lynxchan, purely from a technical perspective, it looks much more professional and mature.

Chans running on jschan:

Known issues

There're some limitations for imageboards running on lynxchan engine (for example, kohlchan.net) due to the lack of support for several features in that engine.

There're some smaller limitations for jschan engine.

There're some very minor limitations for the infinity/OpenIB engine due to the lack of support for several very minor features in that engine.

There're some very minor bugs for 2ch.hk caused by its makaba engine.

Testing

Unit tests:

npm run build
npm test

Real-world tests:

npm run build
npm run test-chan [chan-id]

GitHub

On March 9th, 2020, GitHub, Inc. silently banned my account (erasing all my repos, issues and comments) without any notice or explanation. Because of that, all source codes had to be promptly moved to GitLab. The GitHub repo is now only used as a backup (you can star the repo there too), and the primary repo is now the GitLab one. Issues can be reported in any repo.

License

MIT