Provides natural-language access to PostgreSQL databases, enabling AI agents to query data, explore schemas, generate reports, and analyze business information through the Devart MCP server.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Devart MCP Server for PostgreSQL what are the total sales per region for last quarter?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Devart MCP Server for PostgreSQL enables AI clients to interact with your data through a secure server running in your environment. It turns a regular AI chat into a practical way to work with real-world business data — and it is faster than conventional export or manual querying.
Devart MCP Server for PostgreSQL allows you to:
Work with data intuitively through natural language.
Retrieve the required data for analysis within minutes.
Report on your data faster with AI-powered assistance.
Minimize manual data handling and integration maintenance.
Related MCP server: PostgreSQL MCP Server
Devart MCP Server for PostgreSQL helps AI clients communicate directly with PostgreSQL databases using natural-language prompts. It translates AI requests into structured queries, executes them through Devart connectivity drivers, and returns clean, structured results for seamless AI-powered data access.
To get started with Devart MCP Server for PostgreSQL:
1. Download and install Devart ODBC Driver for PostgreSQL.
2. Download and install Devart MCP Server for PostgreSQL.
3. In Devart MCP Server for PostgreSQL, configure your data connection and integration settings.
4. Run your first natural-language query.
Prerequisites
Before building and running Devart MCP Server for PostgreSQL, ensure the following components are installed:
ADO.NET connection — Devart.AI.McpServer.AdoNet.PostgreSql.csproj Devart dotConnect for PostgreSQL (installed automatically via NuGet during build)
ODBC connection — Devart.AI.McpServer.Odbc.PostgreSql.csproj Devart ODBC Driver for PostgreSQL (requires manual download and installation)
Step 1: Clone the repository
Clone the project repository and navigate to the project directory:
1. Open Command Prompt.
2. Enter the following command:
git clone https://github.com/devart-ai-connectivity/devart-mcp-server-postgresql.git
cd devart-mcp-server-postgresqlStep 2: Build the MCP Server from source
You can build Devart MCP Server for PostgreSQL from source using either of the supported database connectivity technologies: ADO.NET or ODBC.
To build the MCP server with ADO.NET, run the following command:
dotnet publish Devart.AI.McpServer.AdoNet/Devart.AI.McpServer.AdoNet.PostgreSql/Devart.AI.McpServer.AdoNet.PostgreSql.csproj -c ReleasePostgreSql /p:TargetFramework=net8.0The Devart dotConnect for PostgreSQL NuGet package is downloaded and restored automatically.
To build the MCP server with ODBC, select the command based on the bitness of your data source.
For 64-bit data source, run the following command:
dotnet publish Devart.AI.McpServer.Odbc/Devart.AI.McpServer.Odbc.PostgreSql/Devart.AI.McpServer.Odbc.PostgreSql.csproj -c ReleasePostgreSql -r "win-x64" /p:TargetFramework=net8.0For 32-bit data source, run the following command:
dotnet publish Devart.AI.McpServer.Odbc/Devart.AI.McpServer.Odbc.PostgreSql/Devart.AI.McpServer.Odbc.PostgreSql.csproj -c ReleasePostgreSql -r "win-x86" /p:TargetFramework=net8.0Note
The target platform must match the bitness of your ODBC data source.
Step 3: Configure the database connection for the MCP Server
1. Create an mcpserver.json configuration file in the directory containing the built MCP Server executable.
2. In the file, configure the database connection.
Configure a connection with ADO.NET.
Add the following configuration to the mcpserver.json file:
{
"Connections": [
{
"Name": "my_postgresql",
"ConnectionString": "Server=localhost;Port=5432;User Id=your_user;Password=your_password;Database=your_database;License Key=your_license_key;",
"ProtocolType": "stdio"
}
]
}Configure a connection with ODBC.
Add the following configuration to the mcpserver.json file:
{
"Connections": [
{
"Name": "my_postgresql",
"DsnName": "your_dsn_name",
"ProtocolType": "stdio"
}
]
}Configure a connection with ODBC using a connection string.
Add the following configuration to the mcpserver.json file:
{
"Connections": [
{
"Name": "my_postgresql",
"ConnectionString": "Driver={Devart ODBC Driver for PostgreSQL};Server=localhost;User ID=postgres;Password=your_password;Database=your_database;",
"ProtocolType": "stdio"
}
]
}where:
Name — The connection name.
ConnectionString (applies to ADO.NET) — A database-specific connection string used to establish a connection to the target database.
DsnName (applies to ODBC) — The name of your data source.
ProtocolType — A transport protocol. The possible options are: stdio or http.
HttpPort (required if ProtocolType is set to http) — The port number for the http protocol.
Step 4: Run the MCP server
After you configure the MCP Server, you can start it.
Note
This step is required only when
ProtocolTypeis configured ashttp. If you use thestdiotransport protocol, your AI client starts the server automatically.
To start the server with ADO.NET, run the following command:
Devart.AI.McpServer.AdoNet.PostgreSql.exe run my_postgresqlTo start the server with ODBC, run the following command:
Devart.AI.McpServer.Odbc.PostgreSql.exe run my_postgresqlwhere my_postgresql is the name of the ODBC connection.
Step 5: Integrate with Claude Desktop
1. Open claude_desktop_config.json, the Claude configuration file.
Tip
If you can't locate the configuration file, it may not exist yet. To create it, open Claude Desktop and navigate to File > Settings > Developer, then click Edit Config. The folder with the
claude_desktop_config.jsonfile opens.
2. Add one of the following objects, depending on the transport protocol used by MCP Server:
STDIO
{
"mcpServers": {
"devart": {
"command": "C:\\path\\to\\Devart.AI.McpServer.AdoNet.PostgreSql.exe",
"args": [
"run",
"my_postgresql"
]
}
}
}where:
devart is the connector name that will appear in Claude Desktop.
C:\\path\\to\\Devart.AI.McpServer.AdoNet.PostgreSql.exe is the path to the executable file. For an ODBC connection, use Devart.AI.McpServer.Odbc.PostgreSql.exe.
my_postgresql is the connection name specified in the mcpserver.json configuration file.
HTTP
"mcpServers": {
"devart": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"http://localhost:5000/sse"
]
}
}where:
devart is the connector name that will appear in Claude Desktop.
5000 is the MCP Server listening port.
3. Save the file.
4. Restart Claude Desktop.
Devart MCP Server for PostgreSQL is now integrated with Claude, and devart appears in the Claude Desktop app in Customize > Connectors.
You can also integrate Devart MCP Server for PostgreSQL with other AI clients such as Cline, Codex, Cursor, Visual Studio Code, Windsurf, Zed.
Devart MCP Server for PostgreSQL supports integration with the following AI clients:
Claude Desktop
Visual Studio Code
Cursor
Codex
Windsurf
Cline
Zed
...and other MCP-compatible AI clients
Devart MCP Server for PostgreSQL is a practical fit for teams working with PostgreSQL as their primary data source.
Self-service analytics on PostgreSQL data
Let analysts, product managers, sales, finance, and operations teams ask natural-language questions without writing SQL.
Developer copilot for production PostgreSQL databases
Help developers and DBAs investigate real data, inspect schemas, explore joins, and troubleshoot anomalies faster.
Support and customer operations investigation
Let support teams quickly check orders, subscriptions, payments, statuses, and account history stored in PostgreSQL.
Schema exploration and reverse engineering
Use AI to explain tables, relationships, and business logic in legacy or poorly documented PostgreSQL databases.
Ad-hoc reporting and internal analysis
Generate one-off reports, investigate trends, and answer business questions directly from PostgreSQL without building a dashboard first.
Secure on-premises AI access to PostgreSQL
Keep PostgreSQL data inside your infrastructure while giving AI tools a controlled path to work with real business data.
Devart MCP Server for PostgreSQL is distributed as a free single-source MCP server.
To connect to PostgreSQL, the server requires the corresponding Devart ODBC Driver for PostgreSQL, which is a paid product.
A 30-day free trial is available for the Devart ODBC Driver for PostgreSQL.
See the product page and documentation for the latest installation and activation details.
This server cannot be deployed
Query PostgreSQL databases in plain English — LLM-generated, safety-validated SQL.
Deterministic safety, correctness & cost gate that vets Postgres SQL before your AI agent runs it.
Ask questions in plain language, get answers from your business database. No SQL required.
Query your Postgres from ChatGPT or Claude without exposing the database or handing over credentials. Run npx boltschema connect next to your database and it dials out over HTTPS — no inbound firewall rule, no open port, works with localhost and VPC-private databases. Read-only is enforced by a SQL guard, a Postgres READ ONLY transaction, and a scoped role generated for you.
