Forgemax is a Code Mode MCP Gateway that collapses numerous external tools into just two sandboxed interfaces, significantly reducing LLM context window usage and improving performance. It enables Large Language Models to write JavaScript for complex tool chaining, rather than parsing vast JSON schemas, leading to more efficient and secure agentic workflows.

Is forgemax open source?

Yes, licensed under NOASSERTION.

forgemax: Code Mode MCP Gateway that collapses numerous…

Features

01Sandboxed JavaScript Execution: Runs LLM-generated JavaScript in a secure V8 isolate, preventing filesystem or network access, and ensuring fresh runtime per execution.

02LLM Context Window Optimization: Collapses many tools into two standardized MCP tools (`search`, `execute`), drastically reducing token count for tool schemas.

03Advanced Security Measures: Features AST-based code validation, multi-hop alias detection, resource validation, and process isolation to contain LLM-generated code.

04Flexible Tool Orchestration: Supports parallel execution of tool calls, session-scoped data stashing, and dynamic discovery of tools across multiple downstream servers.

05Structured Error Handling: Provides typed `DispatchError` with structured JSON output and fuzzy matching for tool/server name suggestions, enhancing debuggability.

Compatibility

Rust

Runtime

Verified via docs

JavaScript

Code Execution

Verified via docs

Node.js/npm

Installation

Verified via docs

MacOS

Verified via docs

Linux

Verified via docs

Windows

Verified via docs

Quick start

$ npm install -g forgemax

Use cases

↳Building Efficient AI Agents: Create AI agents that interact with a multitude of APIs without overwhelming the LLM's context window with schema details.

↳Secure API Integration for LLMs: Enable LLMs to securely call external APIs by writing JavaScript within a tightly controlled sandbox, preventing unauthorized access or data leakage.

↳Complex Workflow Automation: Orchestrate sophisticated multi-step workflows involving various services and data sources using LLM-generated code.

↳Reducing LLM Inference Costs: Significantly lower token usage for tool descriptions, leading to reduced operational costs for LLM-powered applications.

↳Rapid Tool Prototyping: Quickly integrate and experiment with new tools for LLMs, leveraging Forgemax's abstract interface without deep schema knowledge.

forgemax

Features

Compatibility

Quick start

Use cases

Alternatives

Comments

Deploy on DigitalOcean