MCP Protocol | JSON-RPC 2.0 Implementation | NeuronMCP

Overview

NeuronMCP implements the Model Context Protocol (MCP) with full JSON-RPC 2.0 support, providing a standardized interface for MCP-compatible clients to interact with NeuronDB.

Key Features

JSON-RPC 2.0: Full protocol implementation with stdio, HTTP, and SSE transport
Batch Operations: Transactional batch tool calls for efficient bulk operations
Progress Tracking: Long-running operation progress with progress/get for monitoring
Tool Discovery: Search and filter tools with categorization and metadata
Prompts Protocol: Full prompts/list and prompts/get with template engine support
Sampling/Completions: sampling/createMessage with streaming support for LLM interactions

JSON-RPC 2.0

NeuronMCP implements the JSON-RPC 2.0 specification, providing a standardized protocol for remote procedure calls.

Protocol Features

Request/Response: Standard JSON-RPC request/response format
Error Handling: Comprehensive error codes and messages
Notifications: Support for notification messages
Batching: Batch request support for multiple operations

Transport Modes

NeuronMCP supports multiple transport modes, enabling flexible deployment and integration patterns.

stdio Transport

Standard input/output transport for direct process communication, ideal for local deployments and CLI integration.

HTTP Transport

HTTP-based transport for remote access and web integration, supporting RESTful API patterns.

SSE Transport

Server-Sent Events transport for real-time streaming and progress updates, enabling live monitoring of long-running operations.

Batch Operations

Batch operations enable efficient bulk processing by executing multiple tool calls in a single transaction.

Batch Tool Calls

The tools/call_batch method enables transactional batch execution of multiple tool calls, improving efficiency and ensuring consistency.

Transaction Support

Batch operations are executed within transactions, ensuring atomicity and rollback on failure.

Progress Tracking

Progress tracking enables monitoring of long-running operations through theprogress/get method.

Progress Updates

Long-running operations emit progress updates, providing visibility into operation status and completion.

Progress Monitoring

Clients can poll or subscribe to progress updates, enabling real-time monitoring of operation progress.

Tool Discovery

Tool discovery enables clients to search and filter available tools based on categories, metadata, and capabilities.

Tool Metadata

Each tool includes comprehensive metadata including name, description, parameters, categories, and capabilities.

Tool Filtering

Tools can be filtered by category, capability, or metadata, enabling clients to discover relevant tools efficiently.

Resources & Subscriptions

NeuronMCP provides resources and subscriptions for real-time updates on schema, models, indexes, configuration, workers, and statistics.

Resource Types

Schema Resources: Database schema information with real-time updates
Model Resources: ML model information and status
Index Resources: Vector index configurations and statistics
Config Resources: Server configuration and settings
Worker Resources: Background worker status and metrics
Stats Resources: Database and system statistics with subscriptions

Subscriptions

Resources support real-time subscriptions, enabling clients to receive updates when resources change.