# whodb [Health: Active]

**Category:** 🗄️ Databases  
**Repository:** https://github.com/clidey/whodb  
**GitHub Stars:** 5038  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/whodb

## Description
Self-hosted workspace for browsing and querying databases (Postgres, MySQL, SQLite, MongoDB, Redis and more), with an MCP server for agents.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What whodb does

The whodb MCP server connects AI agents to WhoDB, a self-hosted workspace for browsing and querying databases. WhoDB Community supports PostgreSQL, CockroachDB, YugabyteDB, QuestDB, MySQL, MariaDB, TiDB, SQLite, DuckDB, MongoDB, FerretDB, Redis, Valkey, Dragonfly, Elasticsearch, OpenSearch, ClickHouse, and Memcached. Available operations can vary by database because each source exposes different concepts and capabilities.

WhoDB is designed for inspecting local databases, helping teammates understand unfamiliar data, and working without a full desktop database client. The broader workspace includes a data grid, schema graph, query scratchpad, imports, exports, and mock-data generation. AI providers are optional for the workspace itself and can include hosted or local providers such as OpenAI, Anthropic, Gemini, Ollama, and LM Studio.

## How it works

The MCP server is included with the WhoDB CLI. After installing the CLI, run `whodb mcp serve` to start the server. Agents then use the MCP connection to work with the database access exposed by WhoDB. Connection examples and the complete command reference are maintained in the repository's CLI guide.

WhoDB can run as a Docker container, desktop application, Snap package, or terminal application. When running the workspace in Docker, open its web interface on port 8080 and provide database connection details. The MCP command is a separate CLI entry point, so choose the installation and deployment method that fits the agent environment.

## Setup and configuration

Install the CLI on macOS or Linux with the project's installation script, or install the npm package globally:

```bash
npm install -g @clidey/whodb
whodb mcp serve
```

For a quick WhoDB deployment, run:

```bash
docker run --rm -it -p 8080:8080 clidey/whodb
```

Persistent Docker sessions require a mounted `/data` volume and a saved `WHODB_ENCRYPTION_KEY`. If an HTTPS reverse proxy serves WhoDB, set `WHODB_SECURE=true` so session cookies are sent only over HTTPS. These settings concern the WhoDB deployment rather than a required MCP variable.

## Tools and capabilities

The whodb MCP server is intended for agent access to database exploration and querying through WhoDB. The surrounding WhoDB workspace can:

- Browse, sort, filter, paginate, and edit records in a grid.
- Display table relationships through an interactive schema graph.
- Run queries in a multi-cell scratchpad with autocomplete and history.
- Import and export data.
- Generate mock data for development and testing.
- Provide a terminal workflow through the CLI.

The exact database options available to an individual connection are shown by WhoDB and depend on the selected source.

## Limitations and notes

WhoDB Community is licensed under Apache-2.0 and is intended for self-hosting. Database support is not uniform across all connectors, so an operation available for one source may not be available for another. AI providers are optional and are not required to start the MCP server.

Hosted providers can be configured from the WhoDB interface, while Ollama and LM Studio are available as local options. Provider endpoint variables such as `WHODB_OLLAMA_HOST`, `WHODB_OLLAMA_PORT`, and `WHODB_LMSTUDIO_BASE_URL` are only needed when the default local addresses do not fit the deployment. WhoDB Platform is a separate commercial edition with additional organizational features; the repository describes the Community edition.

_Full upstream README: https://allmcps.com/mcp/whodb/readme_