Nodea
Nodea is an open-source branching AI chat canvas under MIT, built on Anthropic Claude. Every reply is a node on a tree you can pan and zoom. From any node, you fork a new branch, so a conversation grows as a tree instead of one scrolling thread. Model routing picks Haiku, Sonnet or Opus by complexity, and the project ships both as source code and as a hosted product.
What does Nodea do?
Nodea's README calls it "a branching AI chat canvas" and "a non-linear interface for chatting with Anthropic's Claude models." Instead of one long scrolling thread per conversation, every message is stored as a node in a tree. From any node, you fork a new branch, so you can ask the same question a different way, explore an alternative plan, or compare phrasings without overwriting the original path.
The repository is the full source for the hosted product at nodea.ai: the marketing site, the three-panel chat app, the admin dashboard, the Stripe billing integration, Supabase auth, and the Claude streaming endpoints. The chat UI uses XYFlow for pan and zoom over the whole conversation graph, streams tokens through the Vercel AI SDK, routes to Haiku 4.5, Sonnet 4.6 or Opus 4.7 by prompt complexity and plan tier, and supports keyword search and semantic search across every path. An anonymous mode lets a canvas start without a sign-up and claim it later.
Key facts
- Licence: MIT, verified from the LICENSE file.
- Stack, per the README: Next.js 16 App Router, React 19, TypeScript 5, Tailwind CSS v4, Supabase (Postgres with RLS, Supabase Auth), Vercel AI SDK v6 with
@ai-sdk/anthropic, XYFlow for the canvas, Stripe for billing, Vercel for deployment. - Models the README names: Claude Haiku 4.5, Sonnet 4.6 and Opus 4.7.
- Model routing picks the model by prompt complexity and the user's plan tier.
- Live search: keyword search and semantic ("concept") search powered by Claude.
- Color-coded nodes let you tag keepers.
- Hosted plan tiers in the README: Free at $0 (25k daily / 450k monthly tokens), $24 a month Pro for Claude Opus plus 50k daily / 1M monthly tokens.
- Branching from any reply; the data model is a tree of nodes rather than a thread.
- Streaming responses, token by token.
- The README states the hosted product runs on Vercel.
What does it replace, and where does it fall short?
Nodea replaces ChatGPT and Claude.ai for a person who thinks by exploring alternatives: writers, researchers, founders and engineers who are frustrated that "regenerate" and "new chat" each throw away context. It is a candidate on the open-source alternatives to ChatGPT and open-source alternatives to Claude pages.
Where it falls short: Nodea is Claude-only, so a reader who prefers OpenAI, Gemini or an on-device model will find no path in the README. It is a hosted-first product: the self-host route needs Supabase, Stripe, an Anthropic key and a Vercel deployment. For an offline chat UI, Jan or Chatbox. For a Claude-only desktop chat, Claude Desktop MCP if you prefer multiple windows.
How does Nodea run?
Clone the repository, set the Supabase, Anthropic and Stripe keys named in the README, and deploy on Vercel or run npm run dev locally. The three-panel chat app is at /app; the admin dashboard is at /admin. The marketing site runs from the same codebase.
For a reader who does not want to set that up, nodea.ai is the hosted product the repository powers, with a free tier and a $24 a month Pro plan.
Who is Nodea for?
A Claude user who writes or researches, wants to compare two prompts side by side without losing either, and sees a conversation as a tree. For a reader who wants local models or a non-Claude backend, Nodea is not the fit.
What limits does the README state?
- Anthropic Claude only. The README names Haiku 4.5, Sonnet 4.6 and Opus 4.7, with no OpenAI or Gemini provider.
- The hosted free tier is capped at 25k daily and 450k monthly tokens.
- The repository "is the source for the hosted Nodea product"; the paid version uses the same codebase.
- Deployment target in the README is Vercel; other hosts are not described.
Questions people ask
Is Nodea open source?
Yes. The LICENSE file is MIT, verified on the date on this page. The README says you can "fork it, self-host it, or read it to understand the architecture." MIT is an OSI-approved licence with no added conditions.
Does Nodea support models other than Claude?
Not per the README. The project is Claude-only, built on @ai-sdk/anthropic from the Vercel AI SDK v6. Changing the model layer would be a code change rather than a config switch.
How does Nodea compare with ChatGPT or Claude.ai?
The README's own table says a linear chat is a list, Nodea is a tree. On ChatGPT or Claude.ai, trying a different answer means editing the message (which destroys history) or starting a new chat (which loses context). Nodea lets you branch from any node, keep the original path and compare two answers side by side on one canvas. The open-source alternatives to ChatGPT and open-source alternatives to Claude pages list more candidates.
Can I self-host Nodea?
Yes. The README lists Supabase (database and auth), Stripe (billing), Anthropic (model) and Vercel (hosting) as the pieces. The repository is the full product, so a self-host build is possible with those dependencies in place.
How much does the hosted Nodea cost?
Per the README: Free is $0 with 25k daily and 450k monthly tokens; Pro is $24 a month for Claude Opus plus 50k daily and 1M monthly tokens. These are the hosted product's prices; the open-source build has no price.
Sources
- Nodea README and MIT LICENSE: github.com/Elliott-Crosby/Nodea, read 2026-10-19.
More chat interfaces
A self-hosted chat platform for Ollama and OpenAI-compatible models with document search over 9 vector databases, web search, image generation, roles and single sign-on, under a BSD-3 licence with an added branding clause.
LobeHub, formerly LobeChat, is an agent workspace with an agent builder, agent groups, scheduled runs, editable memory and a library of over 10,000 tools and MCP plugins, under a community licence that needs a paid licence to distribute a modified version.
An MIT all-in-one AI app from Mintplex Labs for chatting with your documents with citations, running agents with MCP, and using any of 39 model providers including local llama.cpp and Ollama, as a desktop app or a multi-user Docker server.
An AGPL-3.0 desktop client for Windows, macOS and Linux that connects OpenAI, Gemini, Anthropic and local Ollama or LM Studio models, with 300+ preset assistants, side-by-side model chats, file support and MCP servers.
An AGPL-3.0 desktop app for local models, formerly text-generation-webui, with five engines including llama.cpp and ExLlamaV3, an OpenAI and Anthropic-compatible API, tool calling, file chat, image generation and LoRA training, fully offline.
Added October 19, 2026. Every claim above comes from the project's README, LICENSE or model card, read on October 19, 2026, or from the GitHub API on the date shown in the panel. Found an error? Write to reveneau@licheo.com and it is fixed in the next weekly pass. Repository: github.com/Elliott-Crosby/Nodea.
Want Nodea running on your own servers?
Reveneau sets it up on your hardware, connects it to your product, and builds the software around it.