Supabase
SupabaseManages Supabase projects from an agent: run SQL, inspect tables, apply migrations, and manage edge functions. Powerful enough to be dangerous, so…
Manages Supabase projects from an agent: run SQL, inspect tables, apply migrations, and manage edge functions. Powerful enough to be dangerous, so use a scoped personal access token and a non-production project while you learn its blast radius. Reads Firestore…
Supabase is Postgres with a platform around it; Firebase is a managed document database with tight client SDKs.
The data model decides more than the feature list. Supabase gives an agent SQL over Postgres, which means joins and aggregations an agent can reason about. Firebase gives documents and a query API, which suits client-first apps and suits an agent less well — the same question often needs several queries and a join in application code.
| Option | Install | Source |
|---|---|---|
| Supabase |
npx -y @supabase/mcp-server-supabase
Repository ↗
|
Official |
| Firebase MCP |
npx -y firebase-mcp
Repository ↗
|
Community |
The agent should answer analytical questions with SQL, and the data is relational.
The app is client-first and the data is document-shaped, and the agent is a secondary consumer.
A project going archived, or a maintainer changing the install path, flips this comparison faster than any feature release. Both entries are re-checked weekly, and the date above tells you how much to trust the details.
Manages Supabase projects from an agent: run SQL, inspect tables, apply migrations, and manage edge functions. Powerful enough to be dangerous, so…
Reads Firestore collections and documents and looks up Auth users through a service account. The debugging loop is the use case: why…
Playwright for repeatable tests; Browser MCP for the browser you are already logged into.
Zapier buys connectors; n8n buys control and a fixed bill.
Canva places assets in layouts; Recraft generates editable ones.
Notion is a shared database with a writing surface; Obsidian is local files with links.
Local Docker is one machine's containers; Kubernetes is a cluster with its own permission model.
Dify builds LLM applications; n8n automates systems. The overlap is smaller than the marketing suggests.