惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

S
SegmentFault 最新的问题
G
Google Developers Blog
Stack Overflow Blog
Stack Overflow Blog
WordPress大学
WordPress大学
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
罗磊的独立博客
月光博客
月光博客
IT之家
IT之家
爱范儿
爱范儿
Google DeepMind News
Google DeepMind News
小众软件
小众软件
C
Check Point Blog
B
Blog RSS Feed
H
Help Net Security
博客园 - 司徒正美
L
LangChain Blog
MongoDB | Blog
MongoDB | Blog
B
Blog
The Cloudflare Blog
Apple Machine Learning Research
Apple Machine Learning Research
Microsoft Security Blog
Microsoft Security Blog
M
MIT News - Artificial intelligence
N
Netflix TechBlog - Medium
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知

Hacker News: Show HN

PurrrrrFocus: Pomodoro Timer App - App Store Workflow Engine — Multi-Step Orchestration for Bun RapidPhoto: Pro Photo Editor App - App Store GitHub - DheerG/swarms: Achieve extraordinary results with claude code across a variety of tasks SPICE simulation → oscilloscope → verification with Claude Code — Lucas Gerads Show HN: VCoding – A 5 MB native Windows IDE with no dynamic dependencies Show HN: LLMs don't hallucinate because they're bad at math, it's the format GitHub - Agent-FM/agentfm-core: AgentFM is a peer-to-peer network that turns everyday computers into a decentralized AI supercomputer. AgentFM lets you run massive AI workloads directly across a global mesh of idle CPUs and GPUs. Show HN: Tracking Top US Science Olympiad Alumni over Last 25 Years GitHub - Potarix/agent-hub: One place to talk to all your agents Show HN: Runtime security for AI agents(injection,tool abuse, data exfiltration) GitHub - dubeyKartikay/lazyspotify: Terminal Spotify client for macOS and Linux GitHub - the-banana-tool/king-louie: Easy to use GUI Personal AI Assistant. Win/Linux/Mac. Show HN I made my vacation rental bookable by AI agents–no Airbnb, 0% commission GitHub - basteez/jsf-autoreload: maven plugin to enable hot reload on jsf projects uvm32/hosts/host-gdbstub at main · ringtailsoftware/uvm32 GitHub - labsai/EDDI: Config-driven engine that turns JSON into production-grade AI agents. Multi-agent orchestration, 12+ LLM providers, MCP/A2A protocols, RAG, persistent memory, and enterprise compliance (EU AI Act, GDPR, HIPAA). Built on Quarkus. GitHub - glitchnsec/fortyone-oss: AI Executive Assistant Platform Quickstart | Alien GitHub - muxshed/shed: One stream in, or many. Every destination, simultaneously. No cloud middleman, no per-channel fees, no limits. GitHub - ocrbase-hq/ocrbase: 📄 PDF/IMG ->.MD/JSON Document OCR API for PaddleOCR and GLMOCR. Self-hostable. GitHub - impactjo/home-memory: MCP server that lets your AI assistant remember everything about your home. GitHub - Sets88/dbcls: DbCls is a powerful terminal database client that supports various databases GitHub - neptun2000/heor-agent-mcp GitHub - SeanFDZ/macmind: Single-layer transformer in HyperTalk for the classic Macintosh RollQuation: Math Puzzles - Apps on Google Play GitHub - dropbox/witchcraft Show HN: Agent-cache – Multi-tier LLM/tool/session caching for Valkey and Redis GitHub - opentalon/opentalon: OpenTalon is an open-source platform built from the ground up in Go as a robust alternative to OpenClaw LinkedIn™ 职位抓取工具 - Chrome 应用商店
GitHub - DocSpring/chat_to_map: Use AI to find all your "...
nathan_f77 · 2026-05-01 · via Hacker News: Show HN

ChatToMap

Transform chat exports into geocoded activity suggestions.

License

Overview

ChatToMap extracts "things to do" from WhatsApp, iMessage, Telegram, Facebook Messenger, and LINE exports - restaurants to try, places to visit, trips to take. It finds suggestions buried in years of chat history and puts them on a map.

Features:

  • Parse WhatsApp (iOS/Android), iMessage, Telegram Desktop JSON, Facebook Messenger JSON, and LINE text exports
  • Extract suggestions using multilingual regex patterns, embeddings, and URL detection
  • Classify with AI (activity vs errand, mappable vs general)
  • Scrape metadata from TikTok and YouTube links
  • Geocode locations to coordinates
  • Export to CSV, Excel, JSON, PDF, and interactive HTML map

chat_to_map is the free and open source engine that powers ChatToMap.com. If you are not comfortable with command-line tools or managing API tokens, you can use the ChatToMap web app and pay a small fee to analyze your chat exports.

Installation

# Run directly with npx (zero install)
npx chat-to-map scan "WhatsApp Chat.zip"

# Or install globally
npm install -g chat-to-map

# Or as a library
npm install chat-to-map

CLI Usage

# Free scan - find patterns without API calls
chat-to-map scan <input>

# AI preview - classify top candidates (~$0.01)
chat-to-map preview <input>

# Full analysis with exports
chat-to-map analyze <input>

# Telegram Desktop exports
chat-to-map analyze "/path/to/ChatExport_2026-04-26/result.json"

# Facebook Messenger exports (one chat JSON at a time, extracted from messages.zip)
chat-to-map analyze "/path/to/messages/Friend Name_15.json"

# LINE text exports
chat-to-map analyze "/path/to/[LINE] Chat with Friends.txt"

# List previously processed chats
chat-to-map list

Options

-o, --output-dir <dir>    Output directory (default: ./chat-to-map/output)
-f, --format <formats>    Output formats: csv,excel,json,map,pdf
-r, --region <code>       Region bias for geocoding (e.g., NZ, US, UK)
-n, --limit <num>         Max results for preview/scan
--min-confidence <0-1>    Minimum confidence threshold
--activities-only         Exclude errands (activity_score > 0.5)
--category <cat>          Filter by category
--skip-geocoding          Skip geocoding step
-q, --quiet               Minimal output
-v, --verbose             Verbose output

API Keys

export ANTHROPIC_API_KEY=sk-ant-...   # Required for classification
export GOOGLE_MAPS_API_KEY=AIza...    # Required for geocoding
export OPENAI_API_KEY=sk-...          # Optional for embeddings
export PIXABAY_API_KEY=...            # Optional for images

Images (Opt-In)

Images are not fetched by default. Use --images to enable:

chat-to-map analyze <input> --images

Or fetch images separately for existing results:

chat-to-map fetch-images <input>

Tip: To always fetch images, set it in your config:

chat-to-map config set fetchImages true

Why opt-in?

  • Most external APIs don't allow automated scraping or rate-limit heavily
  • Fetching can take a long time with thousands of activities
  • Google Places photo lookups can be expensive
  • The CLI works perfectly without images

Where images are used:

  • Interactive map HTML (pin details on hover)
  • Activity list view in map export
  • Thumbnails in exported PDFs

Configuration

Configuration is managed in ~/.config/chat-to-map/config.json. homeCountry and timezone are auto-detected from your system and persisted for future runs. Other settings can be set via the config command.

Available settings:

  • homeCountry - Your home country (auto-detected from IP if not set)
  • timezone - Your timezone (auto-detected from system if not set)
  • fetchImages - Whether to fetch images by default
  • cacheDir - Custom cache directory
  • outputDir - Default output directory
  • formats - Default export formats

Manage settings with the config command:

# List all settings
chat-to-map config

# Set a value
chat-to-map config set homeCountry "New Zealand"
chat-to-map config set fetchImages true
chat-to-map config set formats csv,json,map

# Unset a value
chat-to-map config unset cacheDir

Custom config file:

# Use a different config file
chat-to-map analyze <input> --config-file /path/to/config.json

# Or via environment variable
export CHAT_TO_MAP_CONFIG=/path/to/config.json

Library Usage

import {
  parseChat,
  extractCandidatesByHeuristics,
  extractCandidates,  // combined: heuristics + embeddings
  classifyMessages,
  geocodeActivities,
  exportToMapHTML,
  quickScan
} from 'chat-to-map'

// Quick scan (zero API cost)
const scan = quickScan(chatText)
console.log(`Found ${scan.candidates.length} candidates`)

// Parse messages
const messages = parseChat(chatText)

// Extract candidates (heuristics only - sync, free)
const { candidates } = extractCandidatesByHeuristics(messages)

// Or with embeddings (async, requires OpenAI key)
// const result = await extractCandidates(messages, {
//   embeddings: { config: { apiKey: process.env.OPENAI_API_KEY } }
// })
// if (result.ok) candidates = result.value.candidates

// Classify with AI
const result = await classifyMessages(candidates, {
  provider: 'anthropic',
  apiKey: process.env.ANTHROPIC_API_KEY
})

// Geocode and export
if (result.ok) {
  const geocoded = await geocodeSuggestions(
    result.value.filter(s => s.isMappable),
    { apiKey: process.env.GOOGLE_MAPS_API_KEY }
  )
  const html = exportToMapHTML(geocoded)
}

Social Media Scraping

Extract metadata from TikTok and YouTube links found in chats:

import { scrapeUrl, scrapeTikTok, scrapeYouTube } from 'chat-to-map'

// Auto-detect platform
const result = await scrapeUrl('https://youtu.be/abc123')

// Or use platform-specific scrapers
const tiktok = await scrapeTikTok('https://vt.tiktok.com/xxx/')
const youtube = await scrapeYouTube('https://youtube.com/watch?v=xxx')

if (result.ok) {
  console.log(result.metadata.title)
  console.log(result.metadata.description)
  console.log(result.metadata.hashtags)
}

Modules

Module Purpose
parser Parse WhatsApp/iMessage exports
extractor Find candidates via regex and URLs
classifier AI classification (Claude/OpenAI)
embeddings Semantic search with embeddings
geocoder Convert locations to coordinates
scraper Extract metadata from social URLs
export CSV, Excel, JSON, PDF, HTML map
scanner Zero-cost heuristic scanning
cache API response caching

Export Formats

Format Description
CSV All fields, spreadsheet-compatible
Excel Formatted .xlsx with filters
JSON Machine-readable with full metadata
Map Interactive Leaflet.js HTML
PDF Printable report with summary

How to Export Your Chats

Each messaging app has its own export flow. ChatToMap only needs the message text — always pick the "without media" option when offered, both because ChatToMap doesn't use photos/videos and because it keeps the export much smaller.

WhatsApp

iPhone / iPad

  1. Open WhatsApp and tap the chat you want to analyze.
  2. Tap the chat name at the top to open chat info.
  3. Scroll down and tap Export Chat.
  4. Choose Without Media.
  5. Save or share the resulting .zip file.

Android

  1. Open the chat in WhatsApp.
  2. Tap the ⋮ menu in the top-right.
  3. Tap MoreExport chat.
  4. Choose Without media.
  5. Save or share the resulting .zip file.

Desktop (macOS / Windows)

  1. Open WhatsApp Desktop and right-click the chat.
  2. Choose Export chat from the context menu.
  3. Choose Without media.
  4. The .zip file is saved to your Downloads folder.
chat-to-map analyze "WhatsApp Chat - Friends.zip"

iMessage

iMessage has no built-in export. The cleanest path is the open-source imessage-exporter tool, which reads your local iMessage database (or an iPhone backup) and writes one .txt file per conversation that ChatToMap can parse directly.

  1. Install imessage-exporter via Homebrew or Cargo:
    brew install imessage-exporter
    # or
    cargo install imessage-exporter
  2. Grant your terminal Full Disk Access in System Settings → Privacy & Security so it can read ~/Library/Messages/chat.db.
  3. Run the exporter in txt format:
    imessage-exporter -f txt -o ~/imessage_export
    On macOS this reads the local database. To export from an iPhone backup instead:
    imessage-exporter -f txt -p ~/iphone_backup_latest -a iOS -o ~/imessage_export
  4. The output folder contains one .txt file per chat, named after the contact or group. Pass any one of those files to ChatToMap:
chat-to-map analyze "~/imessage_export/Friend Name.txt"

Your messages never leave your computer — both imessage-exporter and chat-to-map run locally.

Telegram

Telegram chat exports are only available from Telegram Desktop. Mobile Telegram apps (iOS / Android) cannot export chat history. On macOS the App Store version is called Telegram Lite — install from https://desktop.telegram.org/, not the older macos.telegram.org build, which doesn't support export.

  1. Open Telegram Desktop and sign in to the account that has the chat.
  2. Open the private chat, group, or channel you want to export.
  3. Use the chat menu → Export chat history.
  4. Choose the machine-readable JSON format when prompted.
  5. Telegram writes a result.json file and any selected media into an export folder. Keep the folder intact or zip it up.
chat-to-map analyze "/path/to/ChatExport_2026-04-26/result.json"
# or a zipped folder
chat-to-map analyze "telegram-export.zip"

Facebook Messenger

Messenger exports messages.zip, which contains every conversation as separate JSON files. Extract it locally and pass one chat at a time.

  1. On a computer, visit https://www.messenger.com/secure_storage/dyi.
  2. Uncheck "Media" — ChatToMap only uses message text and leaving it on makes the export 10× larger.
  3. Pick your date range (e.g. All time).
  4. Click Download file and wait. Messenger needs to assemble the archive — this can take 5–10 minutes for large accounts.
  5. When the file is ready, save messages.zip to your computer.
  6. Double-click messages.zip to extract it. You'll see one .json file per conversation, named like Friend Name_15.json.
chat-to-map analyze "messages/Friend Name_15.json"

LINE

  1. In LINE, open the chat you want to analyze.
  2. Use Settings → Export chat history for that conversation.
  3. Save the resulting .txt file (or the folder/zip if LINE includes media).
  4. Don't edit the exported text before passing it in — the parser reads LINE's original format.
chat-to-map analyze "[LINE] Chat with Friends.txt"

Supported Languages

ChatToMap supports multilingual chats across heuristic extraction, semantic search queries, and AI classification. It can process mixed-language conversations and preserves proper nouns, venues, media titles, and place names in their original script when appropriate.

Current first-class language coverage: Arabic, Bulgarian, Chinese (Simplified), Chinese (Traditional), Czech, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Hungarian, Indonesian, Italian, Japanese, Korean, Norwegian Bokmål, Persian, Polish, Portuguese, Romanian, Russian, Slovak, Spanish, Swedish, Thai, Turkish, Ukrainian, Vietnamese

Privacy

This library and CLI have zero telemetry. We do not collect any data about your usage.

  • No analytics or tracking
  • No phone-home behavior
  • No usage statistics
  • Your chat data stays on your machine

The only network requests are to APIs you explicitly configure (Anthropic/OpenAI for classification, Google Maps for geocoding). These are direct calls to third-party services using your own API keys.

License

AGPL-3.0 - See LICENSE


ChatToMap.com