Forth MCP
Forth MCP is a hosted relay that gives Claude, ChatGPT, and other MCP-speaking AI clients secure access to local, NAT'd, or firewalled MCP servers through a single URL. It is designed to remove the need for port forwarding or a VPN while making local tools usable from anywhere.
At a Glance
- Individual developers and AI power users
- Teams operating agentic workflows
- Enterprises with local or on-premises data and tools
- Organizations needing governed access to MCP servers behind firewalls or NAT
AI Tools by Forth MCP
(1)
Forth MCP
Hosted Relay for Local MCP Servers
Discussions
No discussions yet
Be the first to start a discussion about Forth MCP
Products & Services
A hosted relay plus lightweight connector for exposing local, NAT'd, or firewalled MCP servers to Claude, ChatGPT, and any MCP client through a secure Forth MCP URL and API token. The connector supports local npx packages and plain HTTP endpoints.
A self-hosted deployment that runs the relay on an organization's own network and governs local or remote MCP servers, with unlimited tool calls, SSO support, per-token tool allowlists, usage dashboards, and unlimited audit-log retention.
Market Position
Forth MCP positions itself as an MCP-specific secure relay rather than a general-purpose tunnel: it combines a background connector with MCP endpoint management, per-token tool permissions, quota dashboards, health monitoring, and a self-hosted option. The adjacent competitive set includes ngrok's MCP gateway, Cloudflare Tunnel/Cloudflare Access configurations for securing MCP servers, and open-source client/proxy tools such as mcp-remote and mcp-proxy; Forth's differentiation is the integrated MCP fleet governance and usage controls.
Leadership
Founders
Elad Hefetz
The Hacker News account eladhefetz submitted Forth MCP's public launch and describes its owner as building ForthMCP. A LinkedIn search result identifies Elad Hefetz as CEO of Airfleet, a B2B technology website and marketing company.
Founding Story
Forth MCP was started to solve the gap between cloud-based AI agents and tools/data running on developers' local machines. Its initial vision was a lightweight connector-and-relay architecture that lets remote AI clients call local MCP servers without exposing the machine to the public internet, forwarding ports, or managing VPN infrastructure.
Business Model
Revenue Model
Subscription SaaS with flat monthly plans and hard call caps; paid plans can buy extra call blocks. A self-hosted deployment is offered through a contact-sales model.
Pricing Tiers
1 connector, 500 calls/month, all tools with no filtering, 7-day audit-log retention, and kill switch.
5 connectors, 5,000 calls/month, per-token tool allowlists, 14-day audit-log retention, and optional 5K/20K/50K extra call blocks.
5 connectors, 12,000 calls/month, per-token tool allowlists, 30-day audit-log retention, and optional 5K/20K/50K extra call blocks.
Unlimited connectors, 30,000 calls/month, per-token tool allowlists, 90-day audit-log retention, organization-wide kill switch, and optional extra call blocks.
Unlimited tool calls, local-network deployment, local and remote MCP governance, SSO, per-token tool allowlists, full usage dashboards, and unlimited audit-log retention.
Target Markets
- Individual developers and AI power users
- Teams operating agentic workflows
- Enterprises with local or on-premises data and tools
- Organizations needing governed access to MCP servers behind firewalls or NAT
- Letting cloud-hosted AI agents call tools and access data on a developer's local machine
- Connecting to MCP servers behind NAT or a firewall without port forwarding
- Running multiple local MCP servers for production agentic workflows
- Giving teammates or separate AI clients different tool permissions
- Governing local and remote MCP servers within an enterprise network
- Monitoring MCP tool usage and enforcing call quotas