TrueFoundry's open source AI agent harness TrueForge boasts 30%-75% cheaper task completion than Claude Managed Agents

TrueFoundry, a San Francisco B2B machine-learning startup, released an open-source AI agent harness called TrueForge under the MIT License on August 19, per VentureBeat. The company claims TrueForge can complete enterprise AI tasks at 30%–75% lower cost than Anthropic's Claude Managed Agents.

The cost comparisons come from TrueFoundry's own benchmarking, described in a company blog post. Running TrueForge with the open-source GLM-5.2 model against DevRev's Enterprise-Bench — a suite of multi-step tool-use tasks spanning CRM, issue tracking, and document management — the company says it completed 11 of 14 tasks at $2.90, compared to $11.80 for Claude Managed Agents powered by Claude Opus 4.8, a 75% reduction. Using the same Claude Opus 4.8 model in both harnesses, TrueFoundry claims TrueForge still cost $8.50 versus $11.80 for Claude Managed Agents, roughly a 30% savings. The benchmarks are self-reported; no independent verification was cited.

TrueForge is designed to manage the agent execution loop above TrueFoundry's existing paid AI Gateway product — handling the cycle of reasoning, tool calls, and result processing that an agent repeats until a task is complete. The harness can start locally with a single SQLite process and scale to shared deployments via Docker Compose or Helm with Postgres and Redis. Its cost-reduction architecture centers on context management: it delays loading MCP tool schemas until needed, offloads oversized tool results to files rather than keeping them in the active context window, delegates isolated subtasks to subagents, and compacts long-running conversations at a default threshold of 50,000 tokens. The sandbox is provisioned only when an agent needs to execute code, rather than being held open throughout a run, which TrueFoundry says allows more agents to run concurrently. Co-founder and COO Anuraag Gutgutia told VentureBeat the release responds directly to customer requests: "It is not a replacement. People will use this alongside other harnesses, like the cloud-managed ones or the commercial-provider-managed ones, but this will serve as a way for people to use them in a vendor-neutral way and also at a lower cost."

TrueFoundry was founded in 2021 by Nikunj Bajaj, Abhishek Choudhary, and Gutgutia, who previously worked at Meta and WorldQuant, according to VentureBeat. The company raised a $19 million Series A in February 2025 led by Intel Capital, with participation from Eniac Ventures, Peak XV's Surge, Jump Capital, and angels including Gokul Rajaram and Mohit Aron, bringing total financing to about $21 million per Intel Capital's announcement. As of VentureBeat's January 2026 coverage, the company had more than 30 paid customers, had exceeded $1.5 million in annual recurring revenue, and was processing more than 10 billion requests per month through its AI Gateway. Customers cited by TrueFoundry include NetApp, Siemens Healthineers, ResMed, Automation Anywhere, and Nvidia. In June 2026, the company acquired UK-based MLOps vendor Seldon AI.

The open-source harness does not inherit enterprise access controls on its own — those require TrueFoundry's commercial AI Gateway layer, Gutgutia told VentureBeat. NetApp was a beta user and contributed requirements during development, using the technology for incident response and ticket triage. Automattic was also named as an early user. Gutgutia described the long-term play as making TrueFoundry's gateway the common governance layer beneath whichever agent harnesses — TrueForge or otherwise — enterprises choose to run: "All that traffic should still be flowing through our gateway."