@sandbaseai/cli

@sandbaseai/cli

Give your AI agent superpowers. One command. 2,000+ AI models.

English | 中文 | 日本語 | 한국어 | Español | Français | Deutsch | Português

Your AI coding assistant is smart, but it's trapped in a box. It can't search the web, check social media, generate images, or access real-time data — unless you wire up each API yourself.

SandBase changes that. One command connects your agent to 2,000+ AI models through the Model Context Protocol. No API keys to manage. No configuration headaches.

npx -y @sandbaseai/cli connect

That's it. Your agent now has access to everything.

SandBase Open-Source Stack

• SandBase Harness — self-hosted agent runtime with persistent sessions, sandbox isolation, approvals, audit, and replay.
• SandBase Agent Skills — 88 installable Skills for research, social intelligence, marketing, and business workflows.

See It In Action

After connecting, just ask your agent naturally. It handles the rest.

"Search the web for React 19 new features"

Agent → SandBase Web Search → 10 results in 0.8s
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
1. "React 19 is Here: What's New" — react.dev
2. "React 19 Compiler Deep Dive" — Dan Abramov's blog
3. "Migrating to React 19" — Vercel Engineering
...

"Get trending posts about AI on Twitter"

Agent → SandBase Twitter API → Top 10 posts
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
1. @karpathy: "LLMs are the new CPUs..." (45K ❤️)
2. @emollick: "Agent benchmarks just crossed..." (12K ❤️)
3. @swyx: "The MCP ecosystem is growing fast..." (8K ❤️)
...

"Generate a minimalist logo for my startup 'NightOwl'"

Agent → SandBase Flux Image Gen → 1024x1024 PNG
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
✓ Generated: nightowl-logo.png
  Style: Minimalist, dark theme
  Cost: $0.003
  Time: 2.1s

"Scrape the pricing table from linear.app"

Agent → SandBase Firecrawl Scraper → Structured data
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
✓ Extracted 3 plans:
  Free: $0/mo — Up to 250 issues
  Standard: $8/user/mo — Unlimited issues + cycles
  Plus: $14/user/mo — Advanced features + analytics

"Use GPT-4o to review this pull request for issues"

Agent → SandBase GPT-4o → Analysis complete
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Found 2 issues:
  ⚠️  Race condition in useEffect (line 42)
  ⚠️  Missing error boundary around async component
  ✓  Type safety looks good
  ✓  Test coverage adequate

"Create a 5-second product demo video"

Agent → SandBase Kling Video Gen → Async processing
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⏳ Generating... (45s)
✓ Output: product-demo.mp4 (5s, 720p)
  Cost: $0.10

What's Inside

Supported Clients (17+)

Works with every major AI coding tool:

Preview the complete compatibility catalog without signing in or changing any files:

npx -y @sandbaseai/cli catalog --json

# Connect all detected clients at once
npx -y @sandbaseai/cli connect

# Or target one specific client
npx -y @sandbaseai/cli connect --client cursor

How It Works

┌─────────────────┐         ┌───────────────┐         ┌────────────────────┐
│   Your Agent    │  MCP    │  Local Bridge │  HTTPS  │    SandBase API     │
│  (Cursor, etc.) │────────▶│  (on-demand)  │────────▶│  2,000+ AI models  │
│                 │◀────────│               │◀────────│                    │
└─────────────────┘         └───────────────┘         └────────────────────┘

1. Run connect → CLI opens browser auth → you approve
2. API key saved locally (file permissions 0600)
3. MCP bridge configured for your client
4. Agent calls tools on-demand. No daemon. No background process.

Commands

CLI Commands

sandbase connect [--client <name>]    # Authorize + configure
sandbase doctor [--client <name>]     # Health check
sandbase unregister [--client <name>] # Remove configuration
sandbase catalog --json               # List all supported clients

MCP Tools (available to your agent after connecting)

Workflow your agent follows automatically:

discover → inspect → run

1. sandbase_discover(q: "twitter search")        → finds matching tools
2. sandbase_inspect(name: "twitter_timeline")    → gets schema + pricing
3. sandbase_run(name: "twitter_timeline", ...)   → executes and returns data

Security

• Zero secrets in URLs or CLI args — OAuth device flow with PKCE
• Restricted file permissions — Credentials stored with 0600
• Ownership-aware updates — Existing JSONC comments and user-managed MCP entries are preserved
• Exact rollback — On failure or unregister, only configuration owned by SandBase is removed
• Revoke anytime — One click in the SandBase Dashboard

Get Started

npx -y @sandbaseai/cli connect

Then ask your agent to do something it couldn't before.

Create a free account →

If SandBase saves you setup time, star the repository — it helps more agent users discover the project.

Links

• SandBase Platform
• Documentation
• npm Package
• Dashboard

License

Apache-2.0