Show HN: ThoughtDAG – An editable context graph for LLM conversations
Summary
ThoughtDAG offers an editable context graph for LLM conversations, turning each QA exchange into a node and each edge into part of the model’s context. It provides tools to prune, merge, inspect, and externalize the prompt before generation, with a desktop app and a privacy-focused, self-hosted workflow. The project is open-source (MIT) and supports local-first endpoints for Ollama and OpenAI-compatible setups.