mcp-server-couchbase logo

mcp-server-couchbaseMCP Server to interact with data in Couchbase Clusters

(0)
Daniel NikulshynReviewed by Daniel Nikulshyn·Updated July 2026

Overview

Couchbase MCP Server is a self-hosted server that enables AI agents to interact with data in Couchbase clusters, whether hosted on Capella or self-managed. It provides various tools for cluster health, data schema, key-value operations, queries, and performance, with safety features like read-only mode and fine-grained tool disabling. The server supports STDIO and Streamable HTTP transports and is available as a Python Package Index (PyPI) package and via Docker.

Key features

  • Cluster Health tools
  • Data Schema discovery tools
  • Key-Value operations tools
  • Query and indexing tools
  • Read-only mode and fine-grained tool disabling

Pricing

Model
Free
Category
MCP Servers
Rating
No reviews yet

Use cases

Cluster Setup and Monitoring

Use the Couchbase MCP Server to check cluster credentials, get cluster health status, and list running services.

Data Schema Discovery

Use the server to discover data models and schema in Couchbase clusters, including getting bucket, scope, and collection information.

Key-Value Operations

Perform key-value operations such as getting, upserting, inserting, replacing, and deleting documents by ID in specified scopes and collections.

Pros & Cons

Pros

  • Allows AI agents to connect to and interact with Couchbase clusters
  • Provides safety controls via read-only mode and fine-grained tool disabling
  • Supports both STDIO and Streamable HTTP transports
  • Distributed as a Python Package Index (PyPI) package and via Docker

Cons

  • Some features are disabled by default in read-only mode
  • Requires proper configuration and credentials to connect to Couchbase clusters

Reviews

Sign in to leave a review.

No reviews yet. Be the first!

Q&A

What transport protocols does the server support for AI agents?

The MCP Server works over standard input/output (STDIO) and also offers a streamable HTTP API, letting agents communicate via either local process pipes or HTTP calls.

Asked by Katarzyna Zielinska · Apr 15, 2026

How do I install and run the Couchbase MCP Server?

You can install it as a Python package from PyPI (`pip install couchbase-mcp-server`) or run the Docker image provided in the repository. Both distributions include the server binary and required dependencies.

Asked by Mireille Dupont · Feb 9, 2026

Can I restrict write operations when AI agents use the server?

Yes. Enabling read‑only mode (`CB_MCP_READ_ONLY_MODE=true`) disables write tools such as upsert, insert, and delete by default, allowing only safe read operations.

Asked by Lucas Petit · Feb 9, 2026

Ask a question

MCP Servers alternatives