Forward

Forward MCP Server provides a robust MCP-based interface to Forward Networks and its data, enabling automation, analysis, and integration through a standardized protocol. It exposes a wide range of network tools and supports advanced features like bloom-based filtering and instance locking for reliable deployments.

How to use

You run the server locally and connect an MCP client to its stdio endpoint. The server listens for MCP messages via standard input/output, so you can integrate it with MCP clients that expect a local process you start and control.

Before you start, configure the environment with your Forward Networks credentials and API settings. The server expects your API base URL and API key, along with optional security and performance settings. You can enable bloom filtering for large result sets and configure a persistent bloom index to accelerate repeated queries.

To begin a session, start the server process and then connect your MCP client. The server will handle tool execution, prompts, and resource workflows, and it will respond to supported MCP commands with the appropriate results.

How to install

Prerequisites: You need Go 1.20 or later installed on your system.

Clone the project, build the binary, and verify the executable is ready to run.

git clone https://github.com/forwardnetworks/forward-mcp.git
cd forward-mcp
go build -o forward-mcp ./cmd/server

Run the server

Set the required environment variables before launching the server. These values configure how the server connects to Forward Networks and how it operates.

Required variables include your API base URL and API key. Optional variables tune TLS verification, default network, bloom filtering, and instance locking.

Start the server with the generated binary. It will listen for MCP messages via stdio.

export FORWARD_API_BASE_URL="https://api.forward-networks.example"
export FORWARD_API_KEY="YOUR_API_KEY"
export FORWARD_API_SECRET="YOUR_API_SECRET"
export FORWARD_DEFAULT_NETWORK_ID="default-network-id" # optional
export FORWARD_INSECURE_SKIP_VERIFY=false      # optional

# Optional bloom and lock settings
export FORWARD_BLOOM_ENABLED=true
export FORWARD_BLOOM_THRESHOLD=100
export FORWARD_BLOOM_INDEX_PATH="data/bloom_indexes"
export FORWARD_LOCK_DIR="/tmp"

# Start the server (stdio MCP endpoint)
./forward-mcp

Configuration and behavior

The server reads configuration from environment variables. Key settings include the Forward Networks API base URL and API key, which authenticate your requests. Bloomsearch is available to optimize handling of large results, with automatic bloom filter generation and persistent indexes to speed up repeated queries.

Troubleshooting and notes

If the server cannot start, verify that the API base URL and API key are correct and that network access to Forward Networks is permitted from your environment. Check that the bloom index path exists or is writable if bloom indexing is enabled.

Instance locking helps prevent multiple server instances from running simultaneously. If you enable locking, ensure the lock directory is writable and accessible by the server process.

Additional capabilities

Bloomsearch integration accelerates large result handling by generating bloom filters for NQE results, enabling O(1) membership checks and reducing memory usage. Bloom indexes are persistent, so you can reuse them across sessions and improve overall search performance.

The server exposes a broad set of tools and APIs via MCP, including support for snapshot, location, and device management, as well as prompt workflows and contextual resources.

Notes

This server is designed for easy integration and automation with Forward Networks. It supports both advanced data filtering through bloom techniques and robust instance management to ensure reliable operation in production environments.

Available tools

bloom_search

Bloomsearch integration for large result handling with automatic bloom filter generation and persistent indexes.

snapshot_management

API capabilities to manage snapshots of network data.

location_management

API capabilities to manage locations within the network data.

device_management

API capabilities to manage devices within the network data.

prompts_workflows

Support for prompt-driven workflows and contextual resources.