Skip to main content
Vercel

Vercel

Discover Vercel projects, deployments, and teams.

The Vercel molecule automatically syncs Vercel data into your SixDegree ontology, providing visibility into frontend deployments, build status, and team structure.

  • SixDegree Platform installed
  • Vercel account (personal or team)
  • Vercel API token

discovery:
  enabled: true
  settings:
    # Vercel API credentials
    token: "your_vercel_token"

    # Optional: Vercel API base URL (usually default)
    api_url: "https://api.vercel.com"

    # Optional: Team ID (omit for personal account)
    team_id: "team_xxx"

    # Optional: Project selectors
    selectors:
      - team_id: "team_xxx"
        project_pattern: ".*"  # Regex: all projects

      - team_id: "team_yyy"
        project_pattern: "^frontend-.*"  # Only frontend projects

mcp:
  enabled: true
  settings:
    token: "your_vercel_token"

namespace: "production"

The Vercel molecule discovers:

  • entities.sixdegree.ai/v1/VercelProject - Vercel projects with configuration
  • entities.sixdegree.ai/v1/VercelDeployment - Individual deployments with status
  • entities.sixdegree.ai/v1/VercelTeam - Vercel teams

Each VercelProject entity includes:

  • Name: Project name
  • Framework: Detected framework (Next.js, React, Vue, etc.)
  • Git Repository: Connected repository URL
  • Production Branch: Branch used for production deployments
  • Build Command: Custom build command
  • Output Directory: Build output directory

Each VercelDeployment entity includes:

  • URL: Deployment URL (e.g., project-abc123.vercel.app)

  • State: READY, ERROR, BUILDING, QUEUED, CANCELED

  • Type: PRODUCTION or PREVIEW

  • Git Branch: Source branch

  • Git Commit: Commit SHA

  • Created: Deployment timestamp

  • DEPLOYED_TO - Deployment deployed to Project

  • BELONGS_TO - Project belongs to Team

AI agents can use these Vercel tools:

List all projects in an account or team.

Parameters:

  • team_id (string, optional) - Team ID (omit for personal account)

Example:

User: "What Vercel projects do we have?"

AI: [Calls vercel_list_projects]
    "Found 12 Vercel projects:
     • frontend-web (Next.js) - Production: main branch
     • docs-site (Docusaurus) - Production: master branch
     • marketing-landing (React) - Production: main branch
     • api-gateway (Express) - Production: production branch"

Get detailed project information.

Parameters:

  • project_id (string, required) - Project ID or name
  • team_id (string, optional) - Team ID

Example:

User: "Tell me about the frontend-web project"

AI: [Calls vercel_get_project]
    "frontend-web project:
     • Framework: Next.js 14
     • Repository: github.com/myorg/frontend
     • Production Branch: main
     • Latest Deploy: https://frontend-web.vercel.app
     • Build Command: npm run build
     • Auto deployments: Enabled"

List recent deployments for a project.

Parameters:

  • project_id (string, required) - Project ID or name
  • team_id (string, optional) - Team ID
  • limit (number, optional) - Max deployments to return

Example:

User: "Show recent deployments of frontend-web"

AI: [Calls vercel_list_deployments]
    "Last 5 deployments of frontend-web:
     1. ✅ READY (Production) - 2 hours ago - abc123
        https://frontend-web.vercel.app
     2. ✅ READY (Preview) - 4 hours ago - def456
        https://frontend-web-git-feature-abc.vercel.app
     3. ❌ ERROR (Preview) - 6 hours ago - ghi789
        Build failed: Module not found"

Get detailed deployment information.

Parameters:

  • deployment_id (string, required) - Deployment ID or URL
  • team_id (string, optional) - Team ID

Example:

User: "What happened with deployment abc123?"

AI: [Calls vercel_get_deployment]
    "Deployment abc123:
     • Status: READY ✅
     • Type: Production
     • URL: https://frontend-web.vercel.app
     • Branch: main
     • Commit: abc123def - 'Fix login button'
     • Build Time: 2m 34s
     • Deploy Time: 2 hours ago"

Trigger a new deployment.

Parameters:

  • project_id (string, required) - Project ID or name
  • team_id (string, optional) - Team ID
  • target (string, optional) - Target environment (production or preview)

Example:

User: "Deploy frontend-web to production"

AI: [Calls vercel_trigger_deployment]
    "✅ Deployment triggered for frontend-web
     • Status: BUILDING
     • Preview URL: https://frontend-web-abc123.vercel.app
     • ETA: ~2 minutes"

Get build/runtime logs for a deployment.

Parameters:

  • deployment_id (string, required) - Deployment ID
  • team_id (string, optional) - Team ID

Example:

User: "Why did deployment abc123 fail?"

AI: [Calls vercel_get_deployment_logs]
    "Deployment abc123 logs show:
     • Build started at 14:23:01
     • Installing dependencies... ✅
     • Running build command... ❌
     • Error: Module 'react-icons' not found
     • Suggestion: Run 'npm install react-icons'"

degree discovery run vercel@1.0.0 --config vercel-config.yaml


degree discovery run vercel@1.0.0 --config vercel-config.yaml --dry-run


VERCEL_LOG_LEVEL=debug degree discovery run vercel@1.0.0 --config vercel-config.yaml

Once configured, AI agents can manage Vercel deployments:

You: "What's the status of our frontend deployments?"

AI: [Uses vercel_list_deployments]
    "All frontend projects are healthy:
     • frontend-web: Last deployed 2 hours ago (READY)
     • docs-site: Last deployed 1 day ago (READY)
     • marketing: Last deployed 3 hours ago (READY)"

You: "Deploy the latest changes to production"

AI: [Uses vercel_trigger_deployment]
    "✅ Production deployment started
     Live in ~2 minutes at https://frontend-web.vercel.app"

cd vercel
go build -o ../bin/vercel


cd ..
make dev
  1. Go to Vercel Account Settings → Tokens
  2. Click Create Token
  3. Give it a name (e.g., "SixDegree Discovery")
  4. Select scope:
    • Full Account - For personal projects
    • Specific Teams - For team projects
  5. Copy the token

Required Vercel permissions:

  • Discovery: Read access to projects and deployments
  • MCP Tools:
    • Read: View projects and deployments
    • Deploy: Trigger new deployments
discovery:
  settings:
    token: "your_token"
    # No team_id specified
discovery:
  settings:
    token: "your_token"
    team_id: "team_xxx"

Get team ID from Vercel dashboard URL:

https://vercel.com/team_xxx/settings
                     ^^^^^^^^
                     team_id

Filter projects by name pattern:


selectors:
  - project_pattern: "^frontend-.*"


selectors:
  - project_pattern: "^(?!.*(test|demo|sandbox)).*"
  • Store tokens as secrets in your configuration
  • Use scoped tokens (limit to specific teams if possible)
  • Regularly rotate API tokens
  • Audit token usage in Vercel dashboard
  • Consider separate tokens for discovery and deployments

Authentication failed:

  • Verify token is valid and not expired
  • Check token has access to team
  • Try generating a new token

Projects not found:

  • Verify team_id is correct for team accounts
  • Check token has permission to view projects
  • Ensure projects exist in the account

Deployment failed:

  • Check project build settings in Vercel
  • Verify Git repository is connected
  • Review deployment logs for error details

Rate limiting:

  • Reduce discovery frequency
  • Use project filtering to discover fewer projects
  • Contact Vercel support to increase limits

Link Vercel projects to GitHub repositories:


relations:
  - from: vercel-project:frontend-web
    to: github-repo:myorg/frontend
    type: DEPLOYS_FROM

Link deployments to monitoring dashboards:

relations:
  - from: vercel-deployment:abc123
    to: grafana-dashboard:frontend-metrics
    type: MONITORED_BY

MIT - See LICENSE file in molecules directory