Central Chat - Because that brilliant explanation ChatGPT or Claude gave you three months ago shouldn't be lost forever.

Grab your data archives and get one search bar for everything. Every AI conversation. Zero cloud dependencies. Your conversations are scattered across ChatGPT, Claude, and Raycast—Central Chat brings them together in a single, fast, searchable interface that runs entirely on your machine. Your data never leaves your computer.

Features

• Unified search across all platforms with SQLite FTS5
• Platform theming — Claude (amber), ChatGPT (teal), Raycast (rose)
• Privacy-first — all data stays local in a SQLite database
• Safe re-imports — duplicates are automatically skipped
• Deep linking — shareable URLs for each conversation
• Keyboard-driven — Cmd+K to search, ↑↓ to navigate, ? for help
• Syntax highlighting — code blocks with one-click copy
• Markdown rendering — proper formatting for lists, headings, and more

Quick Start

Prerequisites

• Docker (recommended), OR
• Python 3.10+ and Node.js 18+

Step 1: Get Your Data

Export your chat history from each platform you use:

Platform	How to Export
ChatGPT	chat.openai.com → Settings → Data Controls → Export → Download ZIP
Claude	claude.ai/settings → Export Data → Download ZIP
Raycast	Use raycast-ai-exporter to export your AI chat history
Gemini	Not available yet in Google Takeout

Step 2: Run the App

git clone https://github.com/daveonkels/centralchat.git
cd central-chat
docker compose up -d

Open http://localhost:3000

Step 3: Import Your Data

1. Click Manage Data in the header
2. Upload your exported ZIP or JSON files
3. Click Upload & Import

That's it! Your conversations are now searchable.

Searching

• Type to search across all messages and conversation titles
• Filter by platform using the filter buttons
• Click a result to view the full conversation

Search operators:

• platform:claude — filter by platform (claude, openai, raycast)
• role:user — filter by message role (user, assistant)
• before:2024-01-01 — messages before a date

Keyboard shortcuts: Press ? to see all shortcuts

Screenshots

Tech Stack

Layer	Technology
Frontend	React, TypeScript, Vite
Backend	Python, FastAPI
Database	SQLite with FTS5 full-text search
Deployment	Docker, nginx

Advanced Setup

Manual Setup (without Docker)

Backend

cd backend
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

export DATABASE_PATH=../data/central-chat.db
export IMPORTS_PATH=../imports

uvicorn app.main:app --reload --port 8000

Frontend

cd frontend
npm install
npm run dev

Server Deployment (Traefik)

For deployment to a server with an existing Traefik + Docker Compose setup:

1. Copy project and build images

rsync -avz --exclude 'node_modules' --exclude '.venv' --exclude 'data/*.db' --exclude '.git' \
  ./ yourserver:~/apps/central-chat/

ssh yourserver
cd ~/apps/central-chat
docker build -t central-chat-backend:latest ./backend
docker build -t central-chat-frontend:latest ./frontend

2. Create data directories

mkdir -p ~/data/central-chat/imports

3. Add to your docker-compose.yml

central-chat-backend:
  image: central-chat-backend:latest
  container_name: central-chat-backend
  restart: unless-stopped
  volumes:
    - ~/data/central-chat:/app/data
    - ~/data/central-chat/imports:/app/imports:ro
  environment:
    - DATABASE_PATH=/app/data/central-chat.db
    - IMPORTS_PATH=/app/imports
  networks:
    - proxy

central-chat-frontend:
  image: central-chat-frontend:latest
  container_name: central-chat-frontend
  restart: unless-stopped
  depends_on:
    - central-chat-backend
  networks:
    - proxy
  labels:
    - "traefik.enable=true"
    - "traefik.http.routers.central-chat.rule=Host(`central-chat.example.com`)"
    - "traefik.http.routers.central-chat.entrypoints=https"
    - "traefik.http.routers.central-chat.tls=true"
    - "traefik.http.services.central-chat.loadbalancer.server.port=80"
    - "traefik.docker.network=proxy"

4. Start services

docker compose up -d central-chat-backend central-chat-frontend

Data Storage

• Database: data/central-chat.db
• Imports: imports/ directory
• Safe to delete the database and re-import at any time

Supported Export Formats

Platform	Expected File	Notes
ChatGPT	conversations.json	Full history with message tree structure
Claude	conversations.json	Full history with timestamps
Raycast	raycast_ai_chats.json	Simple format (no per-message timestamps)

API

Endpoint	Description
GET /api/search?q=query	Full-text search
GET /api/conversations	List conversations
GET /api/conversations/{id}	Get conversation with messages
GET /api/conversations/stats	Database statistics
POST /api/import/run	Trigger import

License

MIT