尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
依最近已完成的 8 小時觀測回合間的 GitHub 星號增量排序
Talk to agents. Test trading ideas. Try AI trading agents on real market data — no code required.
翻譯準備中
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權，自行管理介紹，無需管理員核准。
登入並產生驗證檔案，在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案，即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入，將前往登入頁面。登入後請回到此處，再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify，即可驗證所有權並自行管理介紹。
一分鐘完成驗證正在載入完整 README。 以下為已儲存的摘錄。
An evidence-first startup due diligence and investment intelligence platform. Give it a pitch deck, a website and notes; it extracts evidence, analyses eight dimensions, and produces a structured report with an Investment Score. Every statement is labelled as verified or sourced, AI analysis, assumption, or missing, and every claim can be traced to its source.
Status: pre-alpha. The specification set is complete; implementation starts at Phase 0 (see docs/TRACKER.md).檢視完整 README ↗An evidence-first startup due diligence and investment intelligence platform. Give it a pitch deck, a website and notes; it extracts evidence, analyses eight dimensions, and produces a structured report with an Investment Score. Every statement is labelled as verified or sourced, AI analysis, assumption, or missing, and every claim can be traced to its source.
Status: pre-alpha. The specification set is complete; implementation starts at Phase 0 (see docs/TRACKER.md).
ARGUS AI is a research and intelligence tool. It does not provide investment, legal, tax or financial advice, and its outputs are not a substitute for professional due diligence. Verify all material facts independently.
pnpm install
cp .env.example .env.local # fill in the values
pnpm emulators # Auth, Firestore, Storage
pnpm dev # http://localhost:3000
Never commit .env.local or any real documents.
| Script | Purpose |
|---|---|
pnpm dev / pnpm build / pnpm start | Run, build, serve |
pnpm lint / pnpm format / pnpm typecheck | Static checks |
pnpm test / pnpm test:e2e | Unit and integration tests, Playwright |
pnpm check | Lint, typecheck and tests. Must pass before a task is done. |
pnpm emulators | Firebase emulators |
pnpm analyze <path> | Run the analysis engine on a fixture or folder |
pnpm eval | Run the eval harness (calls the live model) |
| Doc | Contents |
|---|---|
docs/PRD.md | Product requirements, epistemic model, report spec |
docs/TRD.md | Architecture, stack, API, security, budgets |
docs/AI_SPEC.md | Evidence contract, pipeline, scoring, verification, prompts, evals |
docs/SCHEMA.md | Zod schemas, Firestore model, rules, canonical fact keys |
docs/APP_FLOW.md | Routes, screens, states, edge cases |
docs/DESIGN.md | Design system |
docs/IMPLEMENTATION_PLAN.md | Phases and kickoff prompts |
docs/TRACKER.md | Task status |
docs/RULES.md | Engineering and AI rules |
docs/PROJECT_MEMORY.md | Decisions, conventions, gotchas, session notes |
CLAUDE.md | Entry point for Claude Code |
src/app routes and API handlers
src/components UI components
src/lib analysis engine, model client, schemas, repositories
evals/ fictional fixtures and eval harness
firebase/ security rules and indexes
docs/ specifications
Start each session from the repository root so CLAUDE.md loads. Use the kickoff prompt for the current phase in docs/IMPLEMENTATION_PLAN.md.
To be decided before any public release.