{"slug": "i-made-thoughtdag-llm-as-an-editable-graph-wires-are-the-context", "title": "I made ThoughtDAG – LLM as an editable graph, wires are the context", "summary": "ThoughtDAG, an open-source tool that turns LLM conversations into editable thought graphs, launched with no install or signup required. The project, hosted on GitHub, lets users wire nodes to control exactly what the model sees, with features like read-only sharing, staleness tracking, and local-first backups. It supports models from Zhipu, Qwen, OpenAI, Anthropic, Google, DeepSeek, Kimi, OpenRouter, and Ollama, with a free tier covering all features.", "body_md": "**Your thinking deserves a map.** An infinite canvas where LLM conversations grow into an editable thought graph.\n\nno install, no signup; the example canvas needs no key\n\n[中文](/chenxiachan/thoughtdag/blob/main/README_ZH.md) · [Quick start](#quick-start) · [More capabilities](#more-capabilities) · [Models](#supported-models) · [Cost & privacy](#cost--privacy)\n\nWires are the context.What the model sees is exactly what wires into the node. Editing the graph edits the model's memory.\n\nOne principle behind every gesture: **the human in the loop, the model on the wires**. No autonomous agent redraws your graph.\n\n|\nThe model sees only what wires in. Delete the noise edge, ask again, and the same prompt returns a clean answer. |\n\n|\nSelect a passage, ask right there. The answer lands on the canvas with its page number, and the p.N chip jumps back to the page. |\n\n|\nMerge nodes into one higher conclusion; weave highlights into a summary. The graph folds inward instead of sprawling. |\n\n|\nHighlights are your judgment, not the model's. Check any subset and weave one passage where every sentence traces back. |\n\n|\nFull cards, takeaway plaques, an icon skeleton: three semantic tiers, every step badged ✕ ⚖ ↩ ?. |\n\n```\n# Online: app.thoughtdag.workers.dev (example canvas needs no key)\n# Local:\nnpm install\nnpm run server    # LLM proxy :3001\nnpm run dev       # → localhost:5173\n# No .env? Connect any OpenAI-compatible endpoint inside the app\n```\n\nThe first launch opens a seeded example canvas: four chapters around one everyday question (why saved articles stay unread), including a reading loop with a real embedded PDF. Environment variables, free keys and configuration details → [docs/setup.md](/chenxiachan/thoughtdag/blob/main/docs/setup.md)\n\n| Capability | What it does |\n|---|---|\n| 📤 Read-only share | One link carries the whole graph: no account, no server storage |\n| 🧭 Staleness & replay | Upstream edits mark the answers they invalidate; replay in dependency order, token estimate first |\n| 🧪 Paradigms | Human-machine workflows saved as files; change the input, replay the experiment |\n| 🔌 Any model | Per-node pins that follow the line; image requests reroute to vision models automatically |\n| 🔒 Local-first | Automatic folder backup writes real files; point it at a synced folder for cross-device |\n\nFull feature list (60+, grouped by area) → [docs/features.md](/chenxiachan/thoughtdag/blob/main/docs/features.md)\n\nZhipu · Qwen · OpenAI · Anthropic · Google · DeepSeek · Kimi · OpenRouter · Ollama, or any OpenAI-compatible endpoint. Requests with images reroute to vision models automatically. Environment variables and default models → [docs/setup.md](/chenxiachan/thoughtdag/blob/main/docs/setup.md)\n\n**The free model tier covers every feature**; a local Ollama runs fully offline** On the hosted demo, model traffic runs browser-direct**: keys never touch the server** PDFs never leave your machine**; only extracted text travels when you ask** The backup format stays backward compatible**; Markdown export is the permanent escape hatch", "url": "https://wpnews.pro/news/i-made-thoughtdag-llm-as-an-editable-graph-wires-are-the-context", "canonical_source": "https://github.com/chenxiachan/thoughtdag", "published_at": "2026-07-22 00:23:12+00:00", "updated_at": "2026-07-22 00:52:21.747475+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-tools", "developer-tools"], "entities": ["ThoughtDAG", "Zhipu", "Qwen", "OpenAI", "Anthropic", "Google", "DeepSeek", "Kimi"], "alternates": {"html": "https://wpnews.pro/news/i-made-thoughtdag-llm-as-an-editable-graph-wires-are-the-context", "markdown": "https://wpnews.pro/news/i-made-thoughtdag-llm-as-an-editable-graph-wires-are-the-context.md", "text": "https://wpnews.pro/news/i-made-thoughtdag-llm-as-an-editable-graph-wires-are-the-context.txt", "jsonld": "https://wpnews.pro/news/i-made-thoughtdag-llm-as-an-editable-graph-wires-are-the-context.jsonld"}}