@linkpress/core

Core library for LinkPress — the building blocks for AI-powered article processing.

What's Inside

| Module | Description | |--------|-------------| | ai | Multi-provider AI summarization & classification (Anthropic, OpenAI, Gemini) | | scraper | URL content extraction with Cheerio | | slack | Slack API client & link extraction utilities |

Installation

npm install @linkpress/core

Quick Start

AI Summarization

import { summarizeArticle } from '@linkpress/core';

const summary = await summarizeArticle(
  'Article Title',
  'Full article content...',
  'https://example.com/article',
  {
    provider: 'anthropic',
    apiKey: process.env.ANTHROPIC_API_KEY,
    model: 'claude-sonnet-4-5-20250929',
    language: 'English',
  }
);

console.log(summary.headline);    // Catchy headline
console.log(summary.tldr);        // 2-3 sentence summary
console.log(summary.keyPoints);   // Array of key points
console.log(summary.tags);        // ['AI', 'TypeScript', ...]
console.log(summary.difficulty);  // 'beginner' | 'intermediate' | 'advanced'

Content Classification

import { classifyContent } from '@linkpress/core';

const classification = await classifyContent(
  'Check out this great article about React hooks!',
  'https://blog.example.com/react-hooks',
  'Understanding React Hooks',
  'A deep dive into useState and useEffect',
  {
    provider: 'anthropic',
    apiKey: process.env.ANTHROPIC_API_KEY,
    model: 'claude-sonnet-4-5-20250929',
  }
);

if (classification.shouldCollect) {
  // Process the article
}

URL Scraping

import { scrapeUrl } from '@linkpress/core';

const content = await scrapeUrl('https://example.com/article');

console.log(content.title);       // Page title
console.log(content.description); // Meta description
console.log(content.content);     // Main text content
console.log(content.image);       // og:image URL
console.log(content.sourceLabel); // 'Blog' | 'GitHub' | 'Newsletter' | ...

Slack Integration

import { SlackClient, extractLinksFromMessages } from '@linkpress/core';

const client = new SlackClient({
  token: 'xoxc-...',
  cookie: 'xoxd-...',
});

// Get user info
const user = await client.getAuthUser();

// List conversations
const conversations = await client.getConversationsList();

// Get messages from a channel
const messages = await client.getConversationHistory('C01234567');

// Extract links from messages
const links = extractLinksFromMessages(messages);
// → [{ url: 'https://...', messageText: '...' }, ...]

Module Imports

Import specific modules for tree-shaking:

// Everything
import { summarizeArticle, scrapeUrl, SlackClient } from '@linkpress/core';

// AI only
import { summarizeArticle, classifyContent, fetchModels } from '@linkpress/core/ai';

// Scraper only
import { scrapeUrl, parseHtml } from '@linkpress/core/scraper';

// Slack only
import { SlackClient, extractLinksFromMessages } from '@linkpress/core/slack';

API Reference

AI Module

summarizeArticle(title, content, url, config): Promise<ArticleSummary>

Generate a magazine-style summary of an article.

Returns:

interface ArticleSummary {
  headline: string;      // Catchy headline
  tldr: string;          // 2-3 sentence summary
  keyPoints: string[];   // 3-5 bullet points
  whyItMatters: string;  // Developer perspective
  keyQuote?: string;     // Memorable quote
  tags: string[];        // Topic tags
  difficulty: 'beginner' | 'intermediate' | 'advanced';
}

classifyContent(messageText, url, title, description, config): Promise<ContentClassification>

Classify content for filtering decisions.

Returns:

interface ContentClassification {
  contentType: 'article' | 'announcement' | 'discussion' | 'reference' | 'social' | 'media' | 'internal' | 'other';
  technicalDepth: 'none' | 'shallow' | 'moderate' | 'deep' | 'expert';
  actionability: 'none' | 'awareness' | 'applicable' | 'reference';
  shouldCollect: boolean;
  reasoning: string;
}

fetchModels(provider, apiKey): Promise<ModelInfo[]>

Fetch available models from an AI provider.

Scraper Module

scrapeUrl(url): Promise<ScrapedContent>

Extract content from a URL.

Returns:

interface ScrapedContent {
  title: string;
  description: string;
  content: string;
  siteName?: string;
  image?: string;
  sourceLabel?: string;  // 'Blog' | 'GitHub' | 'LinkedIn' | 'Newsletter' | 'Article'
}

Slack Module

SlackClient

const client = new SlackClient({ token, cookie });

await client.getAuthUser();                    // Get current user
await client.getConversationsList();           // List all conversations
await client.getConversationHistory(channelId); // Get messages
await client.searchConversations(query);       // Search channels

extractLinksFromMessages(messages): ExtractedLink[]

Extract URLs from Slack messages, filtering out internal tools and non-article links.

Used By

LinkPress CLI — Turn Slack links into a personal tech magazine

Author

Changmin (Chris) Kang