📦 chatpack

Feed your chat history to LLMs. Compress exports 13x with CSV format.

Platforms: Windows • macOS • Linux

The Problem

You want to ask Claude/ChatGPT about your conversations, but:

• Raw exports are 80% metadata noise
• JSON structure wastes tokens on brackets and keys
• Context windows are expensive

The Solution

┌─────────────────┐     ┌──────────┐     ┌─────────────────┐
│ Telegram JSON   │     │          │     │ Clean CSV       │
│ WhatsApp TXT    │ ──▶│ chatpack │ ──▶│ Ready for LLM   │
│ Instagram JSON  │     │          │     │ 13x less tokens │
└─────────────────┘     └──────────┘     └─────────────────┘

Real Numbers

💡 Use CSV for maximum token savings. JSONL is good for RAG pipelines. JSON keeps full structure but wastes tokens.

Use Cases

💬 Chat with your chat history

chatpack tg telegram_export.json -o context.txt

# Paste into ChatGPT: "Based on this conversation, what did we decide about...?"

🔍 Build RAG pipeline

chatpack tg chat.json -f jsonl -t -o dataset.jsonl

# Each line = one document with timestamp for vector DB

📊 Analyze conversations

chatpack wa chat.txt --from "Alice" --after 2024-01-01 -f json

# Filter and export specific messages

Features

• 🚀 Fast — 20K+ messages/sec
• 📱 Multi-platform — Telegram, WhatsApp, Instagram
• 🔀 Smart merge — Consecutive messages from same sender → one entry
• 🎯 Filters — By date, by sender
• 📄 Formats — CSV (13x compression), JSON, JSONL (for RAG)
• 📚 Library — Use as Rust crate in your projects

Installation

Pre-built binaries

Via Cargo

cargo install chatpack

As a library

[dependencies]

chatpack = "0.2"

Quick Start (CLI)

# Telegram

chatpack tg result.json


# WhatsApp  

chatpack wa chat.txt


# Instagram

chatpack ig message_1.json

Output: optimized_chat.csv — ready to paste into ChatGPT/Claude.

Library Usage

Basic example

use chatpack::prelude::*;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Parse a Telegram export
    let parser = create_parser(Source::Telegram);
    let messages = parser.parse("telegram_export.json")?;

    // Merge consecutive messages from the same sender
    let merged = merge_consecutive(messages);

    // Write to JSON
    write_json(&merged, "output.json", &OutputConfig::new())?;

    Ok(())
}

Auto-detect format

use chatpack::parsers::parse_auto;

// Automatically detects Telegram, WhatsApp, or Instagram
let messages = parse_auto("unknown_chat.json")?;

Filter messages

use chatpack::prelude::*;

let parser = create_parser(Source::Telegram);
let messages = parser.parse("chat.json")?;

// Filter by sender
let config = FilterConfig::new()
    .with_user("Alice".to_string());
let alice_only = apply_filters(messages.clone(), &config);

// Filter by date range
let config = FilterConfig::new()
    .after_date("2024-01-01")?
    .before_date("2024-06-01")?;
let filtered = apply_filters(messages, &config);

Output formats

use chatpack::prelude::*;

let messages = vec![
    InternalMessage::new("Alice", "Hello!"),
    InternalMessage::new("Bob", "Hi there!"),
];

// Minimal output (sender + content only)
let config = OutputConfig::new();

// Full metadata (timestamps, IDs, replies, edits)
let config = OutputConfig::all();

// Custom selection
let config = OutputConfig::new()
    .with_timestamps()
    .with_ids();

// Write to different formats
write_json(&messages, "output.json", &config)?;
write_jsonl(&messages, "output.jsonl", &config)?;
write_csv(&messages, "output.csv", &config)?;

Processing statistics

use chatpack::prelude::*;

let original_count = messages.len();
let merged = merge_consecutive(messages);

let stats = ProcessingStats::new(original_count, merged.len());
println!("Compression: {:.1}%", stats.compression_ratio());
println!("Messages saved: {}", stats.messages_saved());

📚 Full API documentation: docs.rs/chatpack

CLI Reference

# Output formats

chatpack tg chat.json -f csv      # 13x compression (default)

chatpack tg chat.json -f json     # Structured array

chatpack tg chat.json -f jsonl    # One JSON per line (for RAG)


# Filters  

chatpack tg chat.json --after 2024-01-01

chatpack tg chat.json --before 2024-06-01

chatpack tg chat.json --from "Alice"


# Metadata

chatpack tg chat.json -t          # Add timestamps

chatpack tg chat.json -r          # Add reply references

chatpack tg chat.json -e          # Add edit timestamps

chatpack tg chat.json --ids       # Add message IDs

chatpack tg chat.json -t -r -e --ids  # All metadata


# Other options

chatpack tg chat.json --no-merge  # Don't merge consecutive messages

chatpack tg chat.json -o out.csv  # Custom output path

Documentation

Supported Platforms

Performance

License

MIT © Mukhammedali Berektassuly