#
ENJA

trueforge

truefoundry/trueforge

TrueForge is an open-source agent harness that provides the runtime layer needed to turn an LLM into a working agent. It supports model calls, MCP tools, sandboxing, approvals, and context management, and offers a chat UI, HTTP API with a TypeScript SDK, and an embeddable UI SDK.

195 2MITUpdated 2026-08-14

Overview

TrueForge is designed to simplify the process of running agents by handling the execution loop, including model calls, tool integration, sandboxing, and session state management. It provides a chat UI, an HTTP API with a TypeScript SDK, and an embeddable UI SDK, making it suitable for both local and hosted environments. It supports multiple model providers, MCP tools, and skills, and can be used in a variety of deployment scenarios.

Capabilities

  • Runs agent execution loops with model calls and tool integration
  • Supports multiple model providers including OpenAI, Anthropic, and Google Gemini
  • Provides a chat UI, HTTP API, and embeddable UI SDK
  • Includes sandboxing for secure code/file execution
  • Supports session persistence, context engineering, and human checkpoints

Best for

TrueForge is ideal for developers and teams looking to build and deploy agents with minimal setup. It is useful for creating chatbots, integrating with external tools, and managing agent workflows in both local and hosted environments.

Works with

CursorGitHub CopilotOpenCode

Security considerations

TrueForge includes sandboxing for secure code execution and stores secrets within the harness, ensuring they are not exposed to external environments. However, local mode should be used with caution as it is not intended for production or internet-facing setups.

Categories

Summary and analysis are original content generated by AI Skills Rank. The skill's source text is not reproduced here — view it on the linked repository.