ctxbridge

Compress any LLM conversation into a single portable context prompt. Paste it into any new session and pick up exactly where you left off.

Supports

• ChatGPT shared links
• Claude shared links
• Gemini (via paste, no shared links yet)
• Any raw text paste

Quick Start

UI (open in browser)

Just open index.html in your browser. No server needed.

CLI

pip install -r requirements.txt

# From a shared URL
python ctxbridge.py --url https://chatgpt.com/share/abc123

# Paste manually
python ctxbridge.py --paste

# From a text file
python ctxbridge.py --file conversation.txt

# Use a different local Ollama model
python ctxbridge.py --url <link> --model qwen2.5-coder:7b

# Skip clipboard copy
python ctxbridge.py --paste --no-copy

How it works

1. Fetches or accepts the conversation
2. Sends it to your local Ollama model (qwen2:7b by default) for compression
3. Outputs a structured [CONTEXT BLOCK] with: Goal, Key Decisions, Important Details, Artifacts, Open Questions
4. Prints to terminal and copies to clipboard

Output format

[CONTEXT BLOCK]

Goal: ...
Key Decisions: ...
Important Details: ...
Artifacts: ...
Open Questions: ...

Paste this at the start of any new ChatGPT, Claude, Gemini, or local LLM session.

Dependencies

• requests, beautifulsoup4, pyperclip, ollama (Python)
• Ollama running locally with at least one model pulled

Built by @rudrabuilds