vercel-services
$
npx mdskill add openai/plugins/vercel-servicesDeploy monorepo backends and frontends together on Vercel.
- Handles Python and Go services in single project deployments.
- Integrates with Vercel CLI and package managers like pip and uv.
- Matches requests containing backend, frontend, or monorepo keywords.
- Executes vercel dev commands to launch multiple services simultaneously.
SKILL.md
.github/skills/vercel-servicesView on GitHub ↗
---
name: vercel-services
description: "Vercel Services — deploy multiple services within a single Vercel project. Use for monorepo layouts or when combining a backend (Python, Go) with a frontend (Next.js, Vite) in one deployment."
metadata:
priority: 7
docs:
- "https://vercel.com/docs/services"
sitemap: "https://vercel.com/sitemap/docs.xml"
pathPatterns:
- 'backend/**'
- 'backend/main.py'
- 'backend/main.go'
- 'backend/go.mod'
- 'backend/pyproject.toml'
- 'backend/requirements.txt'
- 'frontend/**'
- 'apps/*/backend/**'
- 'apps/*/frontend/**'
- 'services/*/vercel.json'
- '*/pyproject.toml'
- '*/go.mod'
bashPatterns:
- '\bvercel\s+dev\b.*-L'
- '\bpip\s+install\b.*fastapi'
- '\buv\s+(sync|pip|run)\b'
- '\bgo\s+(run|build|mod)\b'
- '\bpython\s+-m\s+uvicorn\b'
- '\buvicorn\b'
importPatterns:
- "fastapi"
promptSignals:
phrases:
- "services api"
- "vercel services"
- "multi-service"
- "python backend"
- "go backend"
- "fastapi"
- "deploy backend"
- "backend and frontend"
- "multiple services"
allOf:
- [backend, frontend]
- [python, vercel]
- [go, vercel]
- [backend, deploy]
- [service, monorepo]
- [fastapi, deploy]
anyOf:
- "backend"
- "monorepo"
- "service"
- "python"
- "golang"
noneOf:
- "turborepo cache"
- "turbo.json"
- "aws lambda"
- "docker compose"
minScore: 6
---
# Deploy multi-service projects with Vercel
Services let you deploy multiple independently-built units within a single Vercel project. The typical use case is combining different runtimes (e.g. Python + JavaScript) in one deployment with shared routing and environment variables, but services work for any combination — multiple services of the same runtime, different frameworks, or a mix.
This skill covers **project structure and configuration**. For the actual deployment, defer to the **deployments-cicd** skill.
## How It Works
A service is an independently built unit within your project, deployed to the same domain under a unique subpath. At build time, Vercel builds each service separately. At request time, Vercel routes incoming requests to the correct service based on the URL path prefix (longest prefix wins).
- Services are enabled via the `experimentalServices` field in `vercel.json` (see reference project).
- `vercel dev -L` auto-detects frameworks and runs all services locally as one application, handling routing automatically. The `-L` flag (short for `--local`) runs without authenticating with Vercel Cloud.
- Only `vercel.json` lives at the root. Each service manages its own dependencies independently.
## Configuration
Define services in `vercel.json`:
```json
{
"experimentalServices": {
"web": {
"entrypoint": "apps/web",
"routePrefix": "/"
},
"api": {
"entrypoint": "backend/main.py",
"routePrefix": "/server"
}
}
}
```
The project's Framework Preset must be set to **Services** in the Vercel dashboard.
### Configuration fields
| Field | Required | Description |
|----------------|----------|------------------------------------------------------------|
| `entrypoint` | Yes | Path to the service entrypoint file or directory. |
| `routePrefix` | Yes | URL path prefix for routing (e.g. `/`, `/api`, `/svc/go`). |
| `framework` | No | Framework slug. Pins detection; auto-detected if unset. |
| `memory` | No | Max available RAM in MB (128–10,240). |
| `maxDuration` | No | Execution timeout in seconds (1–900). |
| `includeFiles` | No | Glob patterns for files to include in the deployment. |
| `excludeFiles` | No | Glob patterns for files to exclude from the deployment. |
Do not add unknown fields — they will cause the build to fail.
## Supported runtimes and frameworks
Services is in beta. **Python** and **Go** are tested and production-ready. Other runtimes may work but are not yet validated.
### Python
Works with FastAPI, Flask, Django, or any ASGI/WSGI application. Framework is auto-detected. Set `entrypoint` to the application file (e.g. `"backend/main.py"`). Dependencies go in `pyproject.toml` in the service directory.
### Go
Set `entrypoint` to the service **directory** (e.g. `"backend"`), not a file. **Must** set `"framework": "go"` explicitly in `vercel.json` — auto-detection does not work for Go services. Dependencies in `go.mod` in the service directory.
### Other runtimes (untested)
Vercel supports Node.js, Bun, Rust, Ruby, Wasm, and Edge runtimes for functions. These can theoretically be used as services but are not yet validated.
## Routing
Vercel evaluates route prefixes from longest to shortest (most specific first), with the primary service (`/`) as the catch-all. Vercel automatically mounts services at their `routePrefix`, so service handlers should **not** include the prefix in their routes.
For frontend frameworks mounted on a subpath (not `/`), configure the framework's own base path (e.g. `basePath` in `next.config.js`) to match `routePrefix`.
## Environment variables
Vercel auto-generates URL variables so services can find each other:
| Variable | Example value | Availability | Use case |
|---------------------------------|------------------------------------------|--------------|---------------------------------------|
| `{SERVICENAME}_URL` | `https://your-deploy.vercel.app/svc/api` | Server-side | Server-to-server requests |
| `NEXT_PUBLIC_{SERVICENAME}_URL` | `/svc/api` | Client-side | Browser requests (relative, no CORS) |
`SERVICENAME` is the key name from `experimentalServices`, uppercased. If you define an env var with the same name in project settings, your value takes precedence.
## Usage
1. Read `references/fastapi-vite/` for the canonical project layout.
2. Adapt the structure to the user's chosen runtimes — services can use any supported runtime, not just the ones in the reference.
3. Define service routes **without** the route prefix — Vercel strips the prefix before forwarding.
4. Validate that each service in `vercel.json` has `entrypoint` and `routePrefix`. Only set `framework` when auto-detection fails (required for Go).
## Output
After scaffolding, present the created file structure to the user. After deployment, present the deployment URL (refer to the **deployments-cicd** skill for details).
## Troubleshooting
### 404 on routes after deployment
The project needs the Services framework preset:
1. Go to Project Settings → Build & Deployment → Framework Preset
2. Select **Services** from the dropdown
3. Redeploy
### Routes return unexpected results
1. Ensure all services are picked up by `vercel dev` — check logs. If a service is missing, verify `vercel.json`. Try setting `framework` explicitly.
2. Validate route prefix behavior: handlers declare routes without `routePrefix` (e.g. `/health`), but requests from other services use the full prefix (e.g. `/api/health`).
3. For frontend services on a subpath, confirm the framework's base path config matches `routePrefix`.
More from openai/plugins
- accessibility-and-inclusive-visualizationMake data visualizations accessible and inclusive. Use when the user needs chart or diagram accessibility guidance, text alternatives for complex visuals, color and contrast review, keyboard support, reduced-motion behavior for animation or parallax, or an accessibility QA workflow for exported figures, UML-like diagrams, and dashboards.
- agent-browserBrowser automation CLI for AI agents. Use when the user needs to interact with websites, verify dev server output, test web apps, navigate pages, fill forms, click buttons, take screenshots, extract data, or automate any browser task. Also triggers when a dev server starts so you can verify it visually.
- agent-browser-verifyAutomated browser verification for dev servers. Triggers when a dev server starts to run a visual gut-check with agent-browser — verifies the page loads, checks for console errors, validates key UI elements, and reports pass/fail before continuing.
- agents-sdkBuild AI agents on Cloudflare Workers using the Agents SDK. Load when creating stateful agents, durable workflows, real-time WebSocket apps, scheduled tasks, MCP servers, or chat applications. Covers Agent class, state management, callable RPC, Workflows integration, and React hooks. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.
- ai-elementsAI Elements component library guidance — pre-built React components for AI interfaces built on shadcn/ui. Use when building chat UIs, message displays, tool call rendering, streaming responses, reasoning panels, or any AI-native interface with the AI SDK.
- ai-gatewayVercel AI Gateway expert guidance. Use when configuring model routing, provider failover, cost tracking, or managing multiple AI providers through a unified API.
- ai-generation-persistenceAI generation persistence patterns — unique IDs, addressable URLs, database storage, and cost tracking for every LLM generation
- ai-sdkVercel AI SDK expert guidance. Use when building AI-powered features — chat interfaces, text generation, structured output, tool calling, agents, MCP integration, streaming, embeddings, reranking, image generation, or working with any LLM provider.
- aiq-deploy|
- aiq-research|