Survival LLM
A free AI assistant and survival library (1.5 million offline articles) that runs on your own computer, with no internet. Shows the source of every answer. Mac and Windows.
翻譯準備中
尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
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。
Cross-domain forecasting platform. Next.js 14 (App Router) + TypeScript +
Tailwind, Supabase (Postgres) for data, deployed on Vercel with a daily
cron job pulling fresh indicator data.檢視完整 README ↗Cross-domain forecasting platform. Next.js 14 (App Router) + TypeScript + Tailwind, Supabase (Postgres) for data, deployed on Vercel with a daily cron job pulling fresh indicator data.
entities / indicators — generic schema: any metric, for any country/
city/business/person, at any point in time. See supabase/schema.sql.predictions — the dated, falsifiable public calls that build the CV
track record./api/cron/refresh-indicators — pulls World Bank data server-side and
upserts it. Runs daily via Vercel Cron (vercel.json); can also be hit
manually./api/indicators, /api/predictions — public read-only JSON, cached
at the edge. Ready for the future natural-language interface to query.lib/supabase.ts uses the
service_role key, imported only in Server Components and Route
Handlers) -- nothing reaches the browser bundle.Supabase: open your project's SQL Editor and run, in order:
supabase/schema.sqlsupabase/seed_predictions.sql (optional, seeds the 3 CV predictions)Env vars: cp .env.local.example .env.local and fill in
SUPABASE_URL + SUPABASE_SERVICE_ROLE_KEY from Project Settings ->
API in Supabase (use the service_role key, not anon).
Install Node (this machine didn't have it): the project was set up
using a local copy at ~/.local/node/bin --
export PATH="$HOME/.local/node/bin:$PATH"
Add that line to ~/.zshrc if you want node/npm on PATH in every
new terminal, or just run it once per session.
Install deps and run:
npm install
npm run dev
Open http://localhost:3000.
Load the data (all 217 World Bank countries x 16 indicators, plus a per-country GDP trend projection): with the dev server running, in another terminal:
curl http://localhost:3000/api/cron/refresh-indicators
Then check http://localhost:3000/countries.
Push this repo to GitHub (create an empty repo on github.com, then
git remote add origin <url> && git push -u origin main).
Import the repo in Vercel (vercel.com -> Add New -> Project). No build config needed, Next.js is auto-detected.
In Vercel, add the same env vars as .env.local (Project -> Settings
-> Environment Variables): SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY,
and CRON_SECRET (any random string -- Vercel automatically sends it
as a Bearer token when it triggers the cron job).
Add allforecasts.com as a custom domain in Vercel, then add the
CNAME/A record it gives you into Cloudflare DNS.
The cron in vercel.json runs daily at 06:00 UTC once deployed. To
trigger it manually against production:
curl -H "Authorization: Bearer $CRON_SECRET" https://allforecasts.com/api/cron/refresh-indicators
The schema is deliberately generic so this doesn't require a redesign:
lib/worldbank.ts
(e.g. lib/who.ts), call it from the cron route, tag rows with the
right category/source.entities with
that type; the existing /countries page and CountryTable
component are already generic enough to fork into /cities, etc.