Show HN: ThoughtDAG – An editable context graph for LLM conversations

ThoughtDAG introduces an editable context graph for LLM conversations, addressing the issue of hidden context in linear chat histories. It allows users to visualize, edit, and inspect the context sent to the model, preventing "polluted summaries" by enabling the removal of irrelevant information, such as a "dinner detour." This system ensures reproducible context and cleaner answers by making the "wires" of context visible and editable, contrasting with traditional chat interfaces where the history entering the next request is often invisible.