Your thinking deserves a map. An infinite canvas where LLM conversations grow into an editable thought graph.
中文 · Quick start · How it differs · Research · Models & privacy
Wires are the context. What the model sees is exactly what wires into the node. Editing the graph edits the model's memory.
Many tools put conversations on a canvas. In ThoughtDAG, a wire is not decoration or an execution route. It determines what the model sees next.
One principle behind every gesture: the human in the loop, the model on the wires. No autonomous agent redraws your graph.
 |
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. |
|
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. |
 |
 |
Merge nodes into one higher conclusion; weave your highlights into cited prose. Then zoom through full cards, takeaway plaques and an icon skeleton. The graph folds inward instead of sprawling. |
| Type | What a wire means | Better for |
|---|---|---|
| Linear chat | Conversation history in time order | Quick, simple questions |
| Mind maps and whiteboards | Visual relations for human eyes | Free-form organizing and presenting |
| Branching chat canvases | Parent-child forks of a conversation | Exploring alternative responses |
| Workflow and agent canvases | Data flow or execution order | Automation and orchestration |
| ThoughtDAG | The context the model actually receives next | Deliberate forking, merging, pruning and tracing of long-running thinking |
If you already keep a hand-maintained decision tree in a markdown file, ThoughtDAG is that tree made operational: the model reads exactly the branches you wire in.
Download, open, think. The download page detects your platform and gives you the right installer; Releases keeps every build. macOS builds are signed and notarized. Windows builds are not signed yet and may show a SmartScreen warning.
npm install
npm run server # LLM proxy :3001
npm run dev # → localhost:5173
# No .env? Connect any OpenAI-compatible endpoint inside the appEnvironment variables, local models and connection details → docs/setup.md
Want a ten-second look before installing anything? The hosted demo runs in the browser, and the example canvas needs no key. It is a feature subset: keyless web search, some direct-connection tools and the subscription bridge are desktop/local-only.
ThoughtDAG is also a testbed for a concrete question: when misleading context enters an LLM conversation, how much of the affected path must be removed before the answer recovers?
In the first pilot, deleting only the source repaired 68/72 derailed model-cases. Removing the contaminated subgraph repaired 72/72. This does not explain hidden model reasoning or rank models; it tests how changing visible context changes the next answer.
🧪 Read the first case study · 📊 Methodology and results · 💬 Suggest a model for the next run
| 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
Automatic folder backup keeps the canvas as a live .thoughtdag.json file in your project; Markdown export turns any context chain or selection into a plain .md. Coding agents can read either without a plugin, API or server.
Connect a local Ollama or any OpenAI-compatible endpoint. Built-in presets, subscription connections and environment variables are documented in setup.
- The free model tier covers every feature; a local Ollama runs fully offline
- In the desktop app everything lives on your machine: canvases, keys, documents; on the web demo, model traffic runs browser-direct and 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
With gratitude to @andreilaiter, ThoughtDAG's first supporter, and to everyone helping this independent open-source project grow.
