TrueFoundry’s TrueForge is an MIT-licensed alternative to Claude Managed Agents designed to cut AI agent operational costs by 50%.
On 19 August 2026, enterprise AI infrastructure platform TrueFoundry launched TrueForge, an open-source, vendor-neutral AI agent harness distributed under the MIT License on GitHub and PyPI. Built as a cost-effective alternative to proprietary solutions like Claude Managed Agents, the tool claims to reduce total agent operating expenses by 50%.
TrueForge features vendor neutrality, enabling developers to integrate any LLM and dynamically swap models on a per-task basis to optimise for latency, quality, or expense. Out of the box, the framework includes over 40 built-in tools, sandboxed execution, human-in-the-loop approvals, automatic context compaction, generative UI streaming, and Tavily-powered web search.
For security and governance, all model and Model Context Protocol (MCP) interactions flow through the platform’s gateways, which enforce rate limits, budget caps, role-based access control, and auditing. Alongside the open-source release, TrueFoundry introduced a managed, pay-per-usage hosted version. Early production adopters include NetApp, Automatiq, and TrueFoundry’s internal system, Ask TFY.
TrueFoundry CEO and Co-Founder Nikunj Bajaj explained that TrueForge bridges the gap between ease of use and infrastructure control. Drawing on his experience at Meta, Bajaj noted that the open-source platform allows non-Big-Tech engineering teams to build sophisticated AI agents on any model with enterprise-level security, without the trade-offs or heavy price tag of proprietary solutions.
