The one rule

Wires are the context. What the model sees is exactly what wires into the node. Editing the graph edits the model's memory.

In action

One principle behind every gesture: the human in the loop, the model on the wires. No autonomous agent redraws your graph.

✂️ Delete one edge, get a different answer The model sees only what wires in. Delete the noise edge, ask again, and the same prompt returns a clean answer. Reproduce it in chapter ③ of the example canvas.

📖 Read a paper into a map Select 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. Finish the paper, and the map is drawn.

💎 Thinking condenses in your hands Merge nodes into one higher conclusion; weave highlights into a summary. The graph folds inward instead of sprawling. The human refines in the loop.

🖍️ The passages you marked, woven into cited prose Highlights are your judgment, not the model's. Check any subset and weave one passage where every sentence traces back.

🗺️ Zoom out: thinking becomes a map Full cards, takeaway plaques, an icon skeleton: three semantic tiers, every step badged ✕ ⚖ ↩ ?. The detours are part of the map.

Quick start

# Online: app.thoughtdag.workers.dev (example canvas needs no key)
# Local:
npm install
npm run server    # LLM proxy :3001
npm run dev       # → localhost:5173
# No .env? Connect any OpenAI-compatible endpoint inside the app

The landing page offers the seeded example canvas one labeled click away: 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

Desktop app (beta)

The same app in its own window, with the local server bundled. Grab it from the download page: it detects your platform and hands you the file (macOS Apple Silicon & Intel, Windows, Linux). Every version lives on Releases. No Node, no terminal, and the app tells you in place when a new version ships. Beta builds are not yet code-signed: on macOS right-click → Open on first launch; on Windows choose "More info → Run anyway".

More capabilities

Capability	What it does
📤 Read-only share	One link carries the whole graph: no account, no server storage
🧭 Staleness & replay	Upstream edits mark the answers they invalidate; replay in dependency order, token estimate first
✂️ Clipping	Select a passage or drag a rectangle in the reader; it becomes canvas material with page provenance
🔌 Any model	Per-node pins that follow the line; text-only models read images through their companion text
🔒 Local-first	Automatic folder backup writes real files; point it at a synced folder for cross-device

Full feature list (60+, grouped by area) → docs/features.md

Models & subscriptions

Zhipu · Qwen · OpenAI · Anthropic · Google · DeepSeek · Kimi · OpenRouter · Ollama, or any OpenAI-compatible endpoint. Text-only models read already-indexed images through their companion text; unread images go to a vision model, announced. Environment variables and default models → docs/setup.md

Already paying for a subscription? It plugs in. A ChatGPT plan connects through a one-command local bridge (with ThoughtDAG running locally). GLM Coding and Kimi Code plans issue real API keys: pick the preset, paste the key, done. Setup for all three → docs/setup.md#subscriptions

Cost & privacy

• 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

---