Run the Bidwright Docker server.
One installer brings up the local web app, Fastify API, worker, Postgres, Redis, pgvector, and optional Ollama stack. All product features. No feature gating. No phone home.
macOS / Linux
curl -fsSL https://raw.githubusercontent.com/braedonsaunders/bidwright/main/scripts/launcher/install.sh | bashWindows (PowerShell)
iwr -useb https://raw.githubusercontent.com/braedonsaunders/bidwright/main/scripts/launcher/install.ps1 | iexDocker Desktop
Disk
Memory
Network
What the installer brings up
| Service | Purpose | Port |
|---|---|---|
| web | Next.js frontend (estimating workspace) | 3000 |
| api | Fastify REST API | 3001 |
| worker | BullMQ background jobs (AI, ingestion) | — |
| postgres | Database with pgvector | 5432 |
| redis | Job queue + caching | 6379 |
| ollama | Local embeddings (optional) | 11434 |
Production deployment
For production self-host, run the Docker Compose stack behind a reverse proxy (nginx, Caddy, or Traefik) with TLS. Start simple, then split the database and workers only when usage demands it.
Starting point
4 vCPU / 16 GB RAM / 80 GB SSD
Good for a small estimating team running web, API, worker, Postgres, Redis, and local file storage on one VM.
More headroom
8 vCPU / 32 GB RAM / 160 GB SSD
Use this when package ingest, quote review, or several concurrent users start competing for CPU and memory.
Scale-out move
Managed Postgres + Redis + S3 storage
Move the database, queue, and workspace files off the app VM before adding more web/API/worker instances.