link-reader

Link Reader

Read content from URLs that block direct web fetching by routing through proxy APIs.

Supported Platforms

Twitter / X

URL patterns: twitter.com/*/status/*, x.com/*/status/*, twitter.com/*/article/*, x.com/*/article/*

Proxy: FxTwitter API (api.fxtwitter.com)

Steps:

1. Extract the path after the domain (e.g., /trq212/status/1234567890)
2. Fetch https://api.fxtwitter.com{path} using WebFetch
3. Parse the JSON response
4. Format based on content type (see below)

Regular Tweets

The response JSON has shape: { tweet: { text, author, likes, retweets, ... } }

Format as:

**@{author.screen_name}** ({author.name})

{tweet.text}

{if media.photos: list image URLs}
{if media.videos: list video thumbnail URLs}
{if quote: show quoted tweet inline, indented}

Likes: {likes} · Retweets: {retweets} · Views: {views}
{created_at}

Twitter/X Articles

When tweet.article is present, the tweet is an article (long-form post). The article content uses Draft.js block format.

The article object contains:

• title — article title
• cover_image — hero image metadata (use original_img_url)
• content.blocks[] — array of content blocks
• content.entityMap — links and media referenced by blocks

Converting blocks to markdown:

Each block has a type and text:

• unstyled → plain paragraph
• header-one → # heading
• header-two → ## heading
• header-three → ### heading
• blockquote → > quote
• code-block → fenced code block
• unordered-list-item → - list item
• ordered-list-item → 1. list item
• atomic → look up in entityMap for embedded media/links

Resolving entity ranges:

Each block may have entityRanges: [{ key, offset, length }]. Look up content.entityMap[key]:

• If type: "LINK" → wrap the text span in [text](url) using data.url
• If type: "IMAGE" → insert ![](data.src)

Resolving inline styles:

Each block may have inlineStyleRanges: [{ style, offset, length }]: