365 Open Source Plan #002 · Multi-round AI roundtable with history's greatest thinkers
中文 · 繁體中文 · 日本語 · 한국어 · Español · Français · Deutsch · Português · Italiano · Русский · العربية · हिन्दी · Tiếng Việt · ไทย · Türkçe · Indonesia · বাংলা
Put the world's greatest thinkers in a room and let them debate your problem.
Legend Talk is a multi-round AI roundtable — pick 2-10 historical or contemporary figures, throw in a question, and watch them argue across multiple rounds with a moderator synthesizing each round's key disagreements. Socrates questions Munger's assumptions while Nietzsche challenges them both.
Or just enter a topic and let AI assemble the panel — it picks 3-5 thinkers whose perspectives create productive tension.
Also works as a 1-on-1 thinking tool: consult any of 157 thinkers through their unique frameworks, not generic AI roleplay.
Demo: talk.newzone.top
| Home page | Chat view |
|---|---|
 |
 |
Click the Chat button on any character card to start a private conversation. The thinker responds using their unique framework and perspective.
Enter a topic in the input bar at the top of the home page — AI selects 3-5 thinkers whose perspectives create productive tension and starts the discussion immediately. No character picking needed.
Click the + button on 2-10 character cards to add them to your lineup. A floating bar appears at the bottom showing your selections:
- Click a character's avatar in the bar to remove them
- Click Start Discussion to launch the roundtable
- Click Copy lineup link to share this exact lineup as a URL
You can also click 🎲 Random in the top-right to instantly start a roundtable with 5 random thinkers.
The home page shows 6 curated roundtable templates — pre-built lineups with thinkers whose perspectives genuinely conflict (e.g., "AI & Tech": Karpathy vs Ilya vs Feynman vs Taleb vs Paul Graham). One click to start, with 3 recommended topic suggestions per template to help you dive in.
All roundtable modes show recommended topics before you send your first message:
- Template roundtables — 3 curated questions tailored to each template's theme (available in all 18 languages)
- Manual roundtables — 1 question drawn from each selected character (max 5), so you see a range of perspectives. Topics update as you add or remove characters.
- Moderator — after each round, an AI moderator synthesizes the discussion: identifies the core disagreement, tracks what shifted, and poses a targeted question for the next round
- Stop — cancel generation mid-stream (preserves content already generated)
- Add/remove participants anytime via the participants bar — turning a 1-on-1 into a roundtable or vice versa
- Set rounds — configure how many rounds the thinkers should debate before pausing
- Continue — add more rounds after a discussion completes
- Same characters, new chat — start a fresh conversation with the same characters
- Summarize — one-click AI summary extracting core viewpoints and disagreements
- Share chat — generate a shareable URL containing the full conversation
- Export — save as Markdown or JSON, or generate share cards via json2card (configure API endpoint in Settings)
- Import — restore conversations from a previously exported JSON file (in Settings)
- Branch — fork a new conversation from any message, preserving prior context
Start a conversation directly via URL:
- By name:
/#/chat?chars=苏格拉底,孔子or/#/chat?chars=Socrates,Confucius - By ID:
/#/chat?chars=socrates,confucius - By category:
/#/chat?category=philosophy(starts a roundtable with all thinkers in that category, capped at 10) - Single chat:
/#/chat?chars=socrates - Custom names:
/#/chat?chars=Ada Lovelace,Linus Torvalds(unrecognized names auto-create custom characters)
Available categories: philosophy, strategy, business, finance, history, sociology, psychology, science, literature, art, economics, politics, technology, religion, education.
You can also use the Copy lineup link button in the participants bar or the Copy category link button in the category filter to generate these URLs from the UI.
Language routing: Use a language prefix in the URL to set the UI language, e.g. /#/ja/chat, /#/ko/chat?chars=socrates. You can also use ?lang=zh as a query parameter. Supports all 18 languages.
Beyond the usage above:
- 157 preset thinkers across 15 domains, sorted by fame, with free input for any name
- Conversation search — search across all conversations by title, character name, or message content
- Favorite characters — star your most-used thinkers for quick access
- Settings sync — share settings to another device via URL (API keys AES-encrypted)
- Thinking level — configure model thinking depth (off/low/medium/high)
- Custom models — manually enter any model ID
- Custom LLM — connect any OpenAI-compatible API
- Multi-API support — OpenAI, Anthropic, DeepSeek, Volcengine, Alibaba Bailian, SiliconFlow, Groq, OpenRouter
- 18 languages · Dark mode · Responsive · Local-first (IndexedDB + localStorage dual-write, compatible with WeChat and restricted WebViews)
| Provider | Models |
|---|---|
| OpenAI | GPT-5.4, GPT-5.4 Mini/Nano, o4 Mini, o3, GPT-4.1 series |
| Anthropic | Claude Opus 4.6, Claude Sonnet 4.6, Claude Haiku 4.5 |
| DeepSeek | DeepSeek Chat, DeepSeek Reasoner |
| Volcengine | Doubao Seed 2.0 Pro, Doubao 1.5 series, DeepSeek R1/V3 |
| Alibaba Bailian | Qwen 3.5 Plus, Kimi K2.5, GLM-5, MiniMax M2.5, etc. |
| SiliconFlow | DeepSeek V3/R1, Qwen 2.5 series |
| Groq | LLaMA 4 Scout/Maverick, DeepSeek R1 |
| OpenRouter | Any model via OpenRouter catalog |
All providers support custom model IDs. Default: DeepSeek Chat. You can also add any OpenAI-compatible API via the "Custom" provider option.
npm install
npm run devOpen http://localhost:5173, go to Settings, enter your API key, then start chatting. If you hit a CORS error, the app will prompt you to enable a public proxy with one click.
Some API providers block direct browser requests. CORS proxy is configured per provider in Settings — toggle the switch to enable it. A public proxy (https://cors.api2026.workers.dev) is used by default.
To deploy your own, create a Cloudflare Worker with this code:
Worker code
export default {
async fetch(request) {
const url = new URL(request.url);
const targetUrl = url.pathname.slice(1) + url.search;
if (!targetUrl || !targetUrl.startsWith('https://')) {
return new Response('Usage: /https://target-api.com/path', { status: 400 });
}
if (request.method === 'OPTIONS') {
return new Response(null, {
headers: {
'Access-Control-Allow-Origin': '*',
'Access-Control-Allow-Methods': 'GET, POST, PUT, DELETE, OPTIONS',
'Access-Control-Allow-Headers': '*',
'Access-Control-Max-Age': '86400',
},
});
}
const response = await fetch(targetUrl, {
method: request.method,
headers: request.headers,
body: request.body,
});
const newResponse = new Response(response.body, response);
newResponse.headers.set('Access-Control-Allow-Origin', '*');
return newResponse;
},
};src/
adapters/ # LLM API adapters (OpenAI, Anthropic, etc.)
characters/ # Character presets and custom character generation
components/ # React components
hooks/ # useChat, useRoundtable
i18n/ # Internationalization
stores/ # Zustand state management
utils/ # Prompt building, export, compression, storage
types.ts # Type definitions
React 19, Vite, Tailwind CSS v4, Zustand, i18next, React Router, TypeScript
| Command | Description |
|---|---|
npm run dev |
Start dev server |
npm run build |
Type-check and build for production |
npm run test |
Run tests |
npm run preview |
Preview production build |
Build and deploy the dist/ folder to any static hosting (Vercel, Netlify, GitHub Pages, etc.):
npm run buildUses hash-based routing (/#/chat/..., /#/ja/chat/...) so no server-side routing config needed.
This is project #002 of the 365 Open Source Plan.
One person + AI, 300+ open source projects in a year. Submit your idea →
MIT