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Ask your database for a dashboard. Open-source, local-first AI dashboards — runs on your machine with your own API key.
翻譯準備中
取自儲存庫 README
正在載入完整 README。
檢視完整 README ↗Ask for a dashboard. Keep your data.
Lumyr is an open-source desktop app that turns a sentence into a working Streamlit dashboard. It connects straight to your warehouse, runs every query on your machine, and calls Claude with your own Anthropic API key.
There is no account, no backend, and no Lumyr server. Nothing routes through us, because there is no "us" in the request path.
Landing page · Releases · Contributing · MIT licensed
Being precise about this, because it's the whole point:
| Stays local | Sent to Anthropic under your key |
|---|---|
| Database credentials (OS keychain) | Your prompt |
| Your API key (encrypted via safeStorage) | Table and column names |
| Query execution and results | Rows the agent samples to verify its work |
| The generated dashboard |
No telemetry endpoint, no sign-in, no usage metering. Crash reporting exists but
is off unless you build with a SENTRY_DSN.
The agent is instructed to issue read-only queries. Give it a read-only database role anyway — an enforced permission is worth more than an instruction.
Download a build from Releases, or run from source:
git clone https://github.com/gautamgupta0196/Lumyr
cd Lumyr && npm install
npm run prepare:python # downloads a self-contained Python + dashboard deps
npm run dev
Requirements: Node 20+ and an Anthropic API key from console.anthropic.com.
On first launch, open Settings and paste your API key. It is encrypted with your OS key store and never written into the repo.
BigQuery · PostgreSQL · Snowflake · Redshift · NetSuite · ODBC · CSV/Excel
Generation is handled by the open-source opencode agent,
embedded and run locally. Lumyr gives it a local MCP bridge
(electron/opencode/) so it can list your tables, read schemas, and run
read-only queries while it writes the dashboard — then hands you the resulting
app.py.
npm run dev # Vite dev server + Electron
npm run typecheck # tsc --noEmit
npm run lint # ESLint over src/ and electron/
npm run test:unit # Vitest
npm run build # production build via electron-builder
Architecture notes for contributors live in CLAUDE.md.
MIT — see LICENSE.