Repository avatar
Other Tools
v1.0.1
active

blawby

io.github.Blawby/blawby

Manage and publish restaurant and local business websites through MCP.

Documentation

Blawby

Blawby runs the website, bookings, payments and messaging for restaurants, experience operators, law firms and lodging hosts, managed from the dashboard or from ChatGPT and Claude through one MCP service. Lodging includes homes, nightly availability and prices, direct stay bookings and guest messaging.

Open source under AGPL-3.0.

Nuxt 5 nightly + Nitro 3 + Cloudflare Workers + D1.

Package manager: yarn only. Never npm or pnpm.


Scripts

CommandWhat it does
yarn devNuxt development server with HMR and locally emulated Cloudflare bindings at http://localhost:3000.
yarn local:setupReset local Wrangler state and migrate an empty local D1.
yarn local:seedBuild the synthetic E2E businesses in the running local Worker.
yarn dev:workerBuild and run the production-like Worker locally with Wrangler at http://localhost:3000.
yarn dev:worker:startRun the existing .output Worker build locally without rebuilding it.
yarn buildProduction build → .output/
yarn db:generateGenerate a new migrations/*.sql file from server/db/schema.ts
yarn drizzle:checkVerify server/db/schema.ts hasn't drifted from the live D1 schema
yarn stripe:listenForward Stripe webhooks to localhost (local dev only)
yarn canary:prodProduction-safe authenticated browser canary (read-only checks).
yarn canary:notificationsProduction provider-level email/WhatsApp notification canary.

Local setup

corepack yarn install
cp .env.example .env        # then fill in the application secrets
corepack yarn local:setup   # empty local D1 at the current migration
corepack yarn dev           # http://localhost:3000; leave it running

In a second terminal:

corepack yarn playwright install chromium   # once
corepack yarn local:seed

Use the exact Node version in .nvmrc. .env is the single local configuration for Nuxt, Wrangler, setup scripts and E2E tests; if an older checkout has .dev.vars, move its values into .env and delete it, because Wrangler skips .env when .dev.vars exists.

local:seed builds the same synthetic businesses Playwright uses (tests/e2e/seed) through the app's own boundaries: Better Auth, provisioning, MCP tools and editor uploads. Run local:setup again before reseeding, and after a new migration baseline. No production data reaches a developer machine.

Sign in at /login as platform-admin@playwright.example with E2E_PASSWORD from config/development-auth-fixtures.ts. That admin owns every seeded business:

BusinessDashboardWebsite
Ember & Slice/dashboard/ember-slice-demodemo.localhost:3000
Clayhouse Studio/dashboard/clayhouse-studioclayhouse.localhost:3000
Harbor Legal Aid/dashboard/harbor-legal-aidharbor-legal.localhost:3000

yarn dev is the editing loop. Work on the realtime inbox under yarn dev:worker: Nitro's dev server cannot upgrade the inbox WebSocket. yarn dev:worker:start restarts the last build without rebuilding.

Stop a server by its port, never by process name; several worktrees can run their own server at once:

lsof -ti tcp:3000 | xargs kill

On macOS, raise the file limit if the dev server runs out of handles: ulimit -n 65536.

Yarn accepts new package releases only after seven days. For an urgent reviewed security fix, corepack yarn up <package>@<fixed-version> --no-time-gate and record the advisory in the pull request. Dependency patches live in patches/ and are applied by Yarn's patch: protocol on every install.


Tests

Before pushing:

corepack yarn quality && corepack yarn test:unit && corepack yarn test:d1 && corepack yarn test:migrations && corepack yarn test:mcp

When the MCP tool surface changes, regenerate and commit the catalog:

corepack yarn mcp:catalog:write

corepack yarn chatgpt:submission writes the OpenAI app submission file when you submit to OpenAI; it is not committed.

The E2E suite starts its own Worker on PLAYWRIGHT_PORT (default 3000) with its own local D1, seeds it, and signs in as the fixtures:

PLAYWRIGHT_PORT=3107 corepack yarn test:e2e tests/e2e/<file>.spec.ts --project=chromium

Cloudflare Images credentials exist only in production. Outside it, uploads and social cards that need Cloudflare Images answer 503 by design.


Deployment

Deployment follows the branches in .github/workflows/ci.yml:

  1. Pull requests run Checks and the E2E suite against a local Worker and a local D1 seeded with synthetic businesses. Nothing is deployed for a pull request, and no production data is read.
  2. Pushes to staging run Checks, then deploy the staging Worker, apply staging migrations, and run read-only MCP and tenant rendering against staging itself.
  3. Pushes to main run Checks, then deploy the production Worker, apply production migrations, and run read-only production browser verification.

During an incident, use Cloudflare's deployment history to restore the last known-good production deployment without changing D1 data. Then land the source fix through staging and main.

Production secrets live in Cloudflare → Workers & Pages → krabiclaw → Settings → Variables. CRON_SECRET protects internal scheduled endpoints:

openssl rand -base64 32
yarn wrangler secret put CRON_SECRET

Real-send production canaries (auth OTP, notification email and WhatsApp) run only on demand, from the Production Real-Send Canaries workflow.

Schema and database releases follow release-and-outage-prevention.md.


Schema

server/db/schema.ts is the only schema source of truth; migrations/ holds generated output. Hand-written migrations are prohibited.


Stripe (local)

yarn stripe:listen

Copy the whsec_... signing secret it prints into .env as STRIPE_WEBHOOK_SECRET. Connected accounts use a separate Accounts v2 thin-event destination and signing secret; see Stripe Connect.

No installation packages available.

Remote