87 MCP tools · one endpoint

87 backend tools. One config entry.

Claude, Cursor or any MCP client: your agent scaffolds code, runs Postgres migrations and manages auth — committed to a Git repo you own, with a live preview on every commit.

Start buildingView docs
  1. 1.Create your workspace
  2. 2.Generate an API key
  3. 3.Paste the config — done
mcp-config.json
{
  "mcpServers": {
    "zeeb": {
      "url": "https://backend.zeeb.cloud/mcp/mcp",
      "headers": {
        "Authorization": "Bearer zb_••••••••••••"
      }
    }
  }
}

Paste this into your client. That’s the whole setup.

Works with

ClaudeClaude CodeCursorWindsurfLovableAny MCP client
How it works

From config entry to running backend

Connect once, then let your agent build — five steps to a running live preview in a repo you own.

  1. 01Connect

    Paste one mcpServers entry into Claude, Cursor or Lovable.

    "zeeb": { "url": "…/mcp" }
  2. 02Describe

    Tell your agent what to build, in plain language.

    "Build a shop backend with orders"
  3. 03Agent builds

    It scaffolds code, runs migrations and commits to your Git repo.

    zeeb_generate_crud("shop", "Order")
  4. 04Live preview

    Every commit auto-redeploys its live preview.

    https://9f3a….preview.zeeb.cloud
  5. 05Own & ship

    Push to your own remote and deploy anywhere.

    git push origin main

1 request = 1 tool call = 1 billable unit.

Tool catalog

87 tools. One mcpServers entry.

Everything below is live the moment your key is created — no plugins, no servers to run.

zeeb · mcp surface87 tools
  • zeeb_generate_crud
  • zeeb_create_model
  • zeeb_create_serializer
  • zeeb_create_viewset
  • zeeb_create_route
  • zeeb_create_filterset
  • zeeb_update_viewset
  • zeeb_run_query
  • zeeb_list_tables
  • zeeb_describe_table
  • zeeb_create_migration
  • zeeb_run_migrations
  • zeeb_get_project_reference
  • zeeb_get_preview_url
  • zeeb_redeploy_preview
  • zeeb_get_api_client
  • zeeb_create_user
  • zeeb_list_users
  • zeeb_set_user_password
  • zeeb_create_user_model
  • zeeb_setup_auth
  • zeeb_setup_oauth
  • zeeb_read_file
  • zeeb_write_file
  • zeeb_search_code
  • zeeb_get_project_structure
  • zeeb_set_env
  • zeeb_configure_cors
  • zeeb_read_logs
  • zeeb_run_tests
  • zeeb_generate_dockerfile
  • zeeb_configure_throttling
  • zeeb_configure_versioning
and 50+ more in the docs
Sample callzeeb_generate_crud("blog", "Post")

Scaffolds model, serializer, viewset and route — committed automatically.

Sample callzeeb_run_query("select * from posts")

SQL against your project database, results streamed back.

Sample callzeeb_write_file("blog/views.py", code)

Written into the project repo, committed automatically.

What you can build

A production backend, built by your agent

From an AI frontend’s API layer to a demo-ready MVP — describe it, and your agent ships a backend you own.

The backend for your AI-built frontend

Lovable · v0 · Bolt

Lovable, v0 or Bolt build the UI — Zeeb builds the API, database and auth behind it. Wire them together with one config entry.

Ship an MVP in an afternoon

models · auth · preview

Your agent scaffolds models, CRUD, auth and a database — with a live preview URL to demo the same day. No DevOps.

A real backend, none of the plumbing

git repo · migrations

Skip the dashboards, policies and glue code. You get a Git repo with migrations — code you own, not config trapped in a console.

Platform

Projects, previews, keys and usage

The parts you’d otherwise build around your agent — already there.

  • Git-backed projects you own

    Every project is a real repository. Push to your own remote, move it, leave anytime — the remote stays authoritative.

  • Live preview on every commit

    Each project gets a live preview URL that redeploys on every commit — share work in progress while you build.

    9f3a….preview.zeeb.cloudDEV

  • Teams & API keys

    Isolated workspaces, member roles, and scoped API keys with IP allow-lists and expiry that revoke instantly.

  • Usage analytics

    Live per-tool and per-project usage on your dashboard — you always know where you stand against quota.

  • Managed production hosting

    Coming soon

    Today you deploy your repo to your own infrastructure — you stay in control. One-click managed production hosting is on the roadmap.

portal.zeeb.cloud
Live per-tool and per-project usage on your dashboard
Hosting: IONOS · GermanyJurisdiction: German & EU lawGDPR: AVV / DPA availableExit: git clone
More on our sovereignty
Pricing

Priced per tool call. Never per token.

Start free, upgrade when your agents get busy. No hidden fees.

No credit card required on the Free plan. Compare all plans

FAQ

Frequently asked questions

More in the docs
Which MCP clients are supported?

Anything that speaks MCP over HTTP: Claude (Desktop and Claude Code), Cursor, Windsurf, Lovable, or your own agent. If it can load an mcpServers config, it works with Zeeb.

Can I edit the generated code directly?

Yes. Everything your agent writes is committed to a standard Git repository. Clone it, edit locally, push — your commits and the agent’s tool calls (zeeb_read_file, zeeb_write_file) work on the same repo.

Is the live preview production hosting?

Not yet. The live preview is a development environment that auto-redeploys on every commit — great for building and sharing work in progress. For production you push your repo to your own remote and deploy wherever you like. Managed production hosting is on our roadmap.

Can I point a project at my own Git remote?

Yes — set the Git remote in the project's settings in the portal. From then on your remote stays authoritative; Zeeb clones, commits and pushes against it.

What counts as one tool call?

One call to one tool counts once — zeeb_generate_crud("blog", "Post") is one call, no matter how much code it scaffolds. Tokens are never metered.

Ready to give your agents a backend?

Create your workspace and connect your first MCP client in under five minutes.

Start buildingView docs
Zeeb for developers — 87 MCP tools, one endpoint