CTK: 1,956 OpenAI Conversations as a Long Echo Archive

Following the Long Echo philosophy and the Long Echo Toolkit, I used CTK (Conversation Toolkit) to export and preserve my ChatGPT conversations.

The Archive

The live archive is offline at the moment while I think through what to publish and how. The numbers below describe what was built; if you want a copy of the source corpus, the sanitized data is published as the chatgpt-conversation-corpus repo.

Stats

• 1,956 conversations (sanitized from 2,219, removed 263 containing sensitive data)
• 396 media files (images generated or shared during conversations)
• Self-contained HTML with embedded search
• Dark mode support
• Zero external dependencies

Why Archive Conversations?

AI conversations capture something different from normal documents:

• The iterative process of refining a thought
• Questions that reveal where your understanding has gaps
• Problem-solving and debugging sessions in real time
• Brainstorming that you'd otherwise forget

These conversations are ephemeral by default. Platforms change, accounts get lost, services shut down. Long Echo is about making sure the content outlasts the platform.

CTK Features

CTK follows the same principles as BTK (Bookmark Toolkit):

• Export to HTML -- Self-contained, works offline, opens in any browser
• Automatic sanitization -- Removes conversations containing API keys, passwords, personal identifiers
• Media preservation -- Downloads and embeds images from conversations
• Search -- Full-text search across all conversations
• Multiple views -- Conversation list, timeline, search results

Sanitization

Before export, CTK filters out conversations containing:

• API keys (sk-, ghp_, AKIA)
• Passwords and secrets
• Personal identifiers

This makes the archive safe to share publicly while keeping the intellectual content.

Long Echo in Practice

This archive demonstrates graceful degradation concretely.

The HTML export requires:

• No server
• No JavaScript frameworks
• No external APIs
• No internet connection (after download)

Download this file today, it still works in 20 years. The format is simple enough that future tools can parse and transform it without trouble.

Try CTK

Export your own conversations:

ctk export --db conversations.db --format html output.html --media-dir media

The tool handles parsing conversation JSON, downloading media assets, building searchable indexes, and generating responsive HTML.

---

The archive linked above is a record of thousands of hours of thinking and building with AI assistance. Long Echo is about making sure that record survives the platform it was created on.