Hermes Agent's v0.20.3 Release and $1.5B Funding: A Dual Milestone for Open-Source AI Agents
Two simultaneous signals arrived this week for the open‑source AI agent community. On August 16, 2026, the Hermes Agent team tagged v0.20.3 (v2026.8.16.2), a patch release that rolls up about 125 merged pull requests and solidifies the platform’s move to MCP 2.x and the foundations of bundled Bot Mode. Almost exactly a month earlier, in mid‑July 2026, Nous Research—the company behind Hermes—finalized a $75 million Series B round led by Robot Ventures at a $1.5 billion valuation, with significant participation from Union Square Ventures (USV) and other prominent investors.
Taken together, the two announcements tell a coherent story: Hermes Agent is no longer just a promising side project. It’s a production‑ready agent framework that is now backed by the capital and institutional support needed to keep pushing the boundaries of what open‑source AI agents can do.
What v0.20.3 actually delivers
The release notes for v0.20.3 read like a checklist of platform maturity. The headline feature is the MCP 2.x SDK migration, which aligns Hermes with the latest version of the Model Context Protocol, the standard that lets agents connect to databases, APIs, and tools without hard‑coding integrations. Alongside it comes 2026‑07‑28 stateless protocol support, a behind‑the‑scenes change that makes multi‑turn sessions more resilient and portable.
Perhaps the most user‑visible addition is the bundled Bot Mode (hermes‑bots) plugin, which implements the core “teammate” protocol. This turns Hermes from a single‑assistant tool into a multi‑agent coordinator, letting you spawn specialist bots that can chat, run routines, and message each other—all from the same Hermes instance. For developers who want to script agent behavior programmatically, the new CommandCode provider plugin offers a clean, secure way to inject Python code directly into an agent’s toolchain.
Under the hood, the release tightens security and reliability: subprocess Python runtime ownership hardening isolates agent‑spawned Python environments (PYTHONHOME/PYTHONPATH isolation), and Cua Driver 0.20 runtime contracts bring formal guarantees to the computer‑use tool. The kanban worktree/dispatch fixes eliminate git conflicts when multiple agents edit the same repository, while cron continuity flags and scheduler self‑heal (EMFILE recovery, stale‑claim reconciliation, wedged‑job re‑arm) ensure that scheduled jobs survive crashes and resource exhaustion.
A wave of smaller but no‑less‑important improvements rounds out the list: plugin‑install security scanning, /worktree and /rollback hand‑edit preservation, UTF‑16 file reads for Windows compatibility, and Gemini 3 tool‑call ID preservation that keeps API calls consistent across providers. In short, v0.20.3 is a stability and polish release—the kind that ships when a project is preparing for scale.
The funding side of the equation
While the engineering team was wrapping up those 125 PRs, the business side was closing a deal that doubles the company’s valuation. According to TechCrunch and multiple financial outlets, Nous Research is finalizing a $75 million Series B round at a $1.5 billion valuation, led by Robot Ventures with participation from Union Square Ventures and other investors. The round, expected to close in July 2026, will give the company the runway to expand its team, invest in infrastructure, and continue releasing Hermes Agent as open‑source software.
A $1.5 billion valuation for an open‑source AI agent framework might seem startling, but it reflects a broader trend: developers and enterprises are voting with their feet, choosing Hermes over proprietary alternatives because it gives them full control, no vendor lock‑in, and the ability to run on any LLM provider. The funding validates that choice and ensures that Hermes can keep up with—and often out‑pace—closed‑source competitors.
Why both matter for the ecosystem
Open‑source projects often face a cruel trade‑off: they can either stay lean and nimble, or they can raise money and risk losing their soul. Hermes appears to be threading the needle. The v0.20.3 release proves the technical side is still moving fast, adding features that real users asked for (like the Bot Mode plugin) while hardening the foundation. The funding round, meanwhile, provides the resources to sustain that velocity without turning the project into a walled garden.
For teams building on Hermes, the twin announcements mean they can commit to the platform with confidence. The agent you deploy today will still be supported a year from now, and the roadmap is now backed by a company with the capital to execute it. That’s a rare combination in the often‑volatile world of AI infrastructure.
If you’re looking for a parallel, consider the trajectory of other open‑source tools that started as side projects and grew into foundational infrastructure—projects like Next.js or Tailwind CSS. Hermes Agent is on a similar path, and this week’s news suggests it’s moving faster than anyone expected.
For more on how AI agents are being used in creative companion projects, take a look at TamaBuddy, a sister site exploring AI‑powered digital pets and the social dynamics they create.