LinkedIn/X forces you to compress everything into hot takes. Research papers require months of work. And posting on social media increasingly feels like trying on outfits in ChatGPT until you find one that "fits" – performance over genuine thinking.
I wanted something in between. A place where:
- Half-formed ideas are okay
- Thinking evolves over time (with revision history)
- Garbage gets pruned naturally through format itself and status tracking (exploring → resolved → abandoned)
So I built Kwegg.
How it works:
1. You're in ChatGPT or Claude, working through an idea
2. Say "save this to Kwegg" – it saves your thought with assumptions, evidence, and uncertainties
3. Continue the conversation days later, update your thinking
4. Your thought page shows the full evolution: what you believed, what changed, and why
It uses MCP to connect with ChatGPT and Anthropic Claude. The connector fetches your thoughts and saves updates directly from the conversation. You can even take your ChatGPT thought to Claude and vice versa.
The goal isn't polished content. It's a public record of how thinking actually happens – messy, iterative, and transparent.
All thoughts have:
- Question/Claim – what you're exploring
- Current Position – where you currently stand
- Assumptions – what you're taking for granted
- Evidence – sources and data
- Uncertainties – what you're still unsure about
- Status – exploring, uncertain, resolved, or abandoned
Would love feedback from HN. What's missing? What would make this useful for you?
To use, you can click Learn icon from top, connect with ChatGPT or Claude ( you need paid accounts to add custom MCPs. I have submitted to both and waiting for approval )
LinkedIn/X forces you to compress everything into hot takes. Research papers require months of work. And posting on social media increasingly feels like trying on outfits in ChatGPT until you find one that "fits" – performance over genuine thinking.
I wanted something in between. A place where:
- Half-formed ideas are okay
- Thinking evolves over time (with revision history)
- Garbage gets pruned naturally through format itself and status tracking (exploring → resolved → abandoned)
So I built Kwegg.
How it works:
1. You're in ChatGPT or Claude, working through an idea
2. Say "save this to Kwegg" – it saves your thought with assumptions, evidence, and uncertainties
3. Continue the conversation days later, update your thinking
4. Your thought page shows the full evolution: what you believed, what changed, and why
It uses MCP to connect with ChatGPT and Anthropic Claude. The connector fetches your thoughts and saves updates directly from the conversation. You can even take your ChatGPT thought to Claude and vice versa.
The goal isn't polished content. It's a public record of how thinking actually happens – messy, iterative, and transparent.
All thoughts have:
- Question/Claim – what you're exploring
- Current Position – where you currently stand
- Assumptions – what you're taking for granted
- Evidence – sources and data
- Uncertainties – what you're still unsure about
- Status – exploring, uncertain, resolved, or abandoned
Would love feedback from HN. What's missing? What would make this useful for you?
To use, you can click Learn icon from top, connect with ChatGPT or Claude ( you need paid accounts to add custom MCPs. I have submitted to both and waiting for approval )