Oracle Database MCP Server
# Oracle Database MCP Server
A Model Context Protocol (MCP) server that enables GitHub Copilot and other LLMs to execute read-only SQL queries against Oracle databases.
[](https://www.npmjs.com/package/mcp-oracle-database) [](./LICENSE.md) [](https://glama.ai/mcp/servers/tannerpace/mcp-oracle-database)
---
## ā ļø Breaking Changes in v2.0
`DATE`, `TIMESTAMP`, `TIMESTAMP WITH TIME ZONE`, and `TIMESTAMP WITH LOCAL TIME ZONE` columns are now returned as **formatted strings** (`YYYY-MM-DD HH:mm:ss`) instead of raw JavaScript `Date` objects.
The timezone is controlled by `ORACLE_TIMEZONE` (IANA name). If unset it falls back to the server's system timezone, which may differ across machines.
**Action required ā add `ORACLE_TIMEZONE` to your `.env` or MCP `env` block:**
```env
ORACLE_TIMEZONE=UTC # or e.g. America/Chicago
```
---
## Table of Contents
1. [Installation](#-installation)
2. [Configure VS Code](#-configure-vs-code)
3. [Optional: Create a Read-Only User](#optional-create-a-read-only-user)
4. [Features](#features)
5. [Available Tools](#available-tools)
6. [Configuration Reference](#configuration-reference)
7. [Development](#development)
8. [Security Considerations](#security-considerations)
9. [Troubleshooting](#troubleshooting)
10. [Documentation](#documentation)
11. [Licensing](#licensing)
---
## š¦ Installation
### From npm (quickest)
```bash
npm install -g mcp-oracle-database
```
### Build from Source
```bash
git clone https://github.com/tannerpace/mcp-oracle-database.git
cd mcp-oracle-database
npm install
npm run build
```
> **macOS Apple Silicon (M1/M2/M3/M4) with no Oracle yet?** See the [macOS Setup Guide](./docs/MACOS-SETUP.md) to spin up Oracle XE 21c locally via Colima.
---
## š Configure VS Code
Create `.vscode/mcp.json` in your workspace (or add to your global MCP config).
### Option A ā From Source
```json
{
"servers": {
"oracleDatabase": {
"type": "stdio",
"command": "node",
"args": ["/absolute/path/to/mcp-oracle-database/dist/server.js"],
"env": {
"ORACLE_CONNECTION_STRING": "localhost:1521/XE",
"ORACLE_USER": "system",
"ORACLE_PASSWORD": "OraclePwd123",
"ORACLE_TIMEZONE": "UTC",
"ORACLE_POOL_MIN": "2",
"ORACLE_POOL_MAX": "10",
"QUERY_TIMEOUT_MS": "30000",
"MAX_ROWS_PER_QUERY": "1000",
"ENFORCE_READ_ONLY_QUERIES": "true",
"MCP_MAX_RESPONSE_CHARS": "50000",
"MCP_MAX_ROWS_IN_RESPONSE": "200",
"MCP_MAX_STRING_LENGTH": "500"
}
}
}
}
```
Replace `/absolute/path/to/mcp-oracle-database` with the real path on your machine (e.g. `/Users/yourname/GITHUB/mcp-oracle-database`).
### Option B ā From npm Global Install
```json
{
"servers": {
"oracleDatabase": {
"type": "stdio",
"command": "mcp-database-server",
"env": {
"ORACLE_CONNECTION_STRING": "localhost:1521/XE",
"ORACLE_USER": "your_user",
"ORACLE_PASSWORD": "your_password",
"ORACLE_TIMEZONE": "UTC",
"ORACLE_POOL_MIN": "2",
"ORACLE_POOL_MAX": "10",
"QUERY_TIMEOUT_MS": "30000",
"MAX_ROWS_PER_QUERY": "1000",
"ENFORCE_READ_ONLY_QUERIES": "true",
"MCP_MAX_RESPONSE_CHARS": "50000",
"MCP_MAX_ROWS_IN_RESPONSE": "200",
"MCP_MAX_STRING_LENGTH": "500"
}
}
}
}
```
After saving, reload VS Code and open a Copilot chat in **Agent mode**. Try:
```
"What tables are in the database?"
"Describe the HELP table"
"Show me 5 rows from the HELP table"
```
---
## Optional: Create a Read-Only User
Using `SYSTEM` is fine for local testing. For any real database, use a dedicated read-only user:
```sql
-- Connect: sqlplus system/OraclePwd123@localhost:1521/XEPDB1
CREATE USER readonly_user IDENTIFIED BY secure_password;
GRANT CREATE SESSION TO readonly_user;
GRANT SELECT ANY TABLE TO readonly_user;
-- Or restrict to specific tables:
-- GRANT SELECT ON myschema.orders TO readonly_user;
-- GRANT SELECT ON myschema.customers TO readonly_user;
```
Then update your `.env` or MCP config:
```env
ORACLE_CONNECTION_STRING=localhost:1521/XEPDB1
ORACLE_USER=readonly_user
ORACLE_PASSWORD=secure_password
```
---
## Features
- š **Read-only access** ā Dedicated read-only database user for security
- š” **stdio transport** ā No HTTP server; communicates via standard input/output
- ā” **Connection pooling** ā Efficient Oracle connection management
- š **Schema introspection** ā Query table and column information
- š **Advanced schema discovery** ā 5 specialized tools for tables, relationships, and data patterns
- š¾ **In-memory caching** ā LRU cache with 5-minute TTL for fast repeated access
- š **Audit logging** ā All queries logged with execution metrics
- ā±ļø **Timeout protection** ā Prevents long-running queries
- š”ļø **Result limits** ā Configurable row limits to prevent memory issues
- š **No Oracle Client needed** ā Uses node-oracledb Thin Mode (pure JS, works on Apple Silicon)
### Architecture
```
GitHub Copilot / LLM
ā (MCP Protocol)
MCP Client (spawns process)
ā (JSON-RPC over stdio)
MCP Server (Node.js)
ā (node-oracledb Thin Mode)
Oracle Database (read-only user)
```
---
## Available Tools
### Core Tools
#### `query_database`
Execute read-only SQL SELECT queries.
```json
{
"query": "SELECT table_name FROM user_tables FETCH FIRST 10 ROWS ONLY",
"maxRows": 10
}
```
#### `get_database_schema`
Get a table list or column details for a specific table.
```json
{ "tableName": "ORDERS" }
```
### Schema Discovery Tools
| Tool | Purpose | Cached |
|------|---------|:------:|
| `listTables` | All accessible tables with metadata & optional row counts | ā
|
| `describeTable` | Column types, constraints, primary/foreign keys | ā
|
| `getTableRelations` | Foreign key relationships in JSON | ā
|
| `getSampleValues` | Sample values to understand data formats | ā |
| `suggestRelatedTables` | Find related tables by FK, naming, shared columns | ā |
š See [Schema Discovery Documentation](./docs/SCHEMA-DISCOVERY.md) for full details and examples.
### Example Copilot Prompts
```
"List all tables in the database"
"Describe the ORDERS table and its relationships"
"How many active users are there?"
"What are the top 5 products by sales this month?"
"Show me recent transactions for customer ID 12345"
```
---
## Configuration Reference
All settings can go in `.env` or as `env` keys in your VS Code MCP config.
```env
# Oracle Database Connection
ORACLE_CONNECTION_STRING=localhost:1521/XE # host:port/service
ORACLE_USER=system
ORACLE_PASSWORD=OraclePwd123
# Timezone (REQUIRED for consistent date/timestamp output)
ORACLE_TIMEZONE=UTC # IANA timezone, e.g. America/New_York
# Connection Pool
ORACLE_POOL_MIN=2
ORACLE_POOL_MAX=10
# Query Safety
QUERY_TIMEOUT_MS=30000 # max query time in ms
MAX_ROWS_PER_QUERY=1000 # max rows Oracle will fetch
MAX_QUERY_LENGTH=50000 # max SQL length in chars
ENFORCE_READ_ONLY_QUERIES=true # reject non-SELECT statements
# MCP Response Limits
MCP_MAX_RESPONSE_CHARS=50000 # hard cap on total response size
MCP_MAX_ROWS_IN_RESPONSE=200 # max rows per tool call response
MCP_MAX_STRING_LENGTH=500 # max chars per string field
# Logging
LOG_LEVEL=info
ENABLE_AUDIT_LOGGING=true
ENABLE_FILE_LOGGING=true
LOG_DIR=./logs
NODE_ENV=development
```
> **Large schemas:** If your database has 500+ tables, raise `MCP_MAX_RESPONSE_CHARS` to `100000`.
---
## Development
### Scripts
```bash
npm run build # Compile TypeScript ā dist/
npm run dev # Watch mode compilation
npm run clean # Remove dist/
npm run typecheck # Type-check without compiling
npm start # Start MCP server (requires build first)
npm run test-client # Core tool tests against live Oracle DB
npm run test-discovery # Schema discovery tool tests
```
### Project Structure
```
mcp-oracle-database/
āāā src/
ā āāā server.ts # MCP server entry point
ā āāā client.ts # Core test client
ā āāā test-discovery.ts # Discovery tools test client
ā āāā config.ts # Zod-validated configuration
ā āāā database/
ā ā āāā oracleConnection.ts # Connection pool manager
ā ā āāā queryExecutor.ts # Query execution + safety checks
ā ā āāā types.ts
ā āāā tools/
ā ā āāā queryDatabase.ts # query_database tool
ā ā āāā getSchema.ts # get_database_schema tool
ā ā āāā discovery/ # 5 schema discovery tools + cache
ā āāā utils/
ā āāā logger.ts # Lightweight file + console logger
ā āāā responseFormatter.ts # MCP response size management
āāā dist/ # Compiled output (git-ignored)
āāā .env # Your credentials (git-ignored)
āāā .env.example # Template
āāā package.json
```
---
## Security Considerations
1. **Read-Only User** ā Database user should have only SELECT privileges in production
2. **SQL Safety** ā The server trusts the LLM to generate valid SQL; the read-only user is the safety net
3. **Query Limits** ā Row count and timeout limits prevent resource exhaustion
4. **Audit Logging** ā All queries logged with timestamps for review
5. **Local First** ā Designed to run on your machine; can still connect to remote databases
---
## Troubleshooting
### Connection failed
```
Error: ORA-12545: Connect failed because target host or object does not exist
```
- Is Oracle running? `docker ps | grep oracle-xe`
- Check the port: `docker ps` should show `0.0.0.0:1521->1521/tcp`
- Try `localhost:1521/XE` for SYSTEM, `localhost:1521/XEPDB1` for other users
### Wrong service name
| Service | Use for |
|---------|---------|
| `localhost:1521/XE` | SYSTEM user, DBA operations |
| `localhost:1521/XEPDB1` | Regular application users |
### Permission denied
```
Error: ORA-00942: table or view does not exist
```
```sql
GRANT SELECT ANY TABLE TO your_user;
```
### Response too large
```
Response for tool 'listTables' exceeded MCP_MAX_RESPONSE_CHARS
```
```env
MCP_MAX_RESPONSE_CHARS=100000
```
### Colima / Docker issues (macOS)
See [macOS Setup Guide ā Troubleshooting](./docs/MACOS-SETUP.md#troubleshooting).
### Thin Mode note
This project uses node-oracledb **Thin Mode** ā a pure JavaScript driver that requires no Oracle Instant Client. It works on all platforms including Apple Silicon Macs.
---
## Documentation
š **Guides:**
- [macOS Setup Guide](./docs/MACOS-SETUP.md) ā Local Oracle XE setup on Apple Silicon
- [Schema Discovery Guide](./docs/SCHEMA-DISCOVERY.md) ā Advanced schema introspection tools
- [Schema Discovery Quick Reference](./docs/SCHEMA-DISCOVERY-QUICKREF.md) ā Cheat sheet for all discovery tools
- [Schema Discovery Examples](./docs/SCHEMA-DISCOVERY-EXAMPLES.md) ā MCP message examples
- [VS Code Integration Guide](./docs/VSCODE-INTEGRATION.md) ā Set up with GitHub Copilot
- [Claude Desktop Integration Guide](./docs/CLAUDE-INTEGRATION.md) ā Set up with Claude Desktop
- [MCP Integration Guide](./docs/MCP-INTEGRATION.md) ā MCP protocol deep dive
- [Architecture Overview](./docs/ARCHITECTURE.md) ā System architecture diagram
- [Logging Configuration](./docs/LOGGING.md) ā Logging setup and configuration
š **Custom Instructions:**
- [`.github/copilot-instructions.md`](./.github/copilot-instructions.md) ā Project-wide Copilot instructions
- [`.github/instructions/`](./.github/instructions/) ā Language-specific coding guidelines
<a href="https://glama.ai/mcp/servers/tannerpace/mcp-oracle-database">
<img width="380" height="200" src="https://glama.ai/mcp/servers/tannerpace/mcp-oracle-database/badges/card.svg" alt="Oracle Database MCP server" />
</a>
---
## Licensing
This project is licensed under the [GNU Affero General Public License v3.0 (AGPL-3.0)](./LICENSE.md).
You are free to use, modify, and distribute this software under the terms of the AGPL-3.0. Any modified version deployed as a network service must make its source code available to users of that service.
See [LICENSE.md](./LICENSE.md) for the full license text.
---
> Oracle is a registered trademark of Oracle Corporation.
> This project is not affiliated with, endorsed by, or sponsored by Oracle Corporation.
TDQS
Scored across 2 tools
The two tools have clearly distinct purposes: get_database_schema retrieves metadata about the database structure, while query_database executes SQL queries for data retrieval. There is no overlap or ambiguity between them.
Both tools follow a consistent verb_noun naming pattern (get_database_schema, query_database) with clear, descriptive names that align with their functions. The naming is uniform and predictable.
With only 2 tools, the server feels under-scoped for an Oracle Database MCP Server, as it lacks essential operations like data manipulation (INSERT, UPDATE, DELETE), transaction management, or administrative tasks, making it incomplete for typical database workflows.
The tool surface is severely incomplete for a database server, covering only schema inspection and read-only queries. Missing are critical operations such as data modification, stored procedure execution, user management, and other core database functionalities, leading to significant gaps.