Hermes Summit: The Evolution of a Web3 Team

Original Title: "IOSG Weekly Brief | The Rise of Hermes: An Advanced Web3 Team's Journey #340"
Original Author: Jacob Zhao, IOSG Ventures

The phenomenal growth of Hermes did not stem from exclusive technology based on the OpenClaw principle that cannot be replicated. Instead, it precisely closed a key window of opportunity in the formation of the Personal Agent category, establishing a "Challenger Growth System" that takes over the already educated user base of OpenClaw and creates a more authentic experience difference based on "Delegation Trust" compared to a "self-evolution" narrative. As professional executive Agents become stronger, users still need a long-term online steward worthy of trust.

Opening the public app rankings of OpenRouter, the Hermes Agent leads the platform with a Token usage of 30.5 trillion, ranking first in Productivity, Coding Agents, Personal Agents, and CLI Agents categories, far ahead of well-known Agents such as OpenClaw and Claude Code.

▲ Figure 1: Historical Snapshot of Hermes Agent on OpenRouter (as of August 4, 2026, dynamic page data will change over time)

Although OpenRouter's statistics methodology cannot cover the industry-wide Token consumption of direct official APIs (such as Claude or Codex native subscriptions), as the world's largest AI large-scale model routing and aggregation platform, its rankings carry a strong "benchmark" significance. While in the high-end professional task domain, the core business workflows of many users—complex code generation, architecture design, high-value data analysis—still flow towards Claude Code and ChatGPT, Hermes maintains an advantage in backend automation, message entry response, long-term online monitoring, and lightweight task scheduling use cases. As an Agent product built by a Web 3 team, Hermes has achieved far beyond expected dissemination, community, and usage intensity success, prompting us to pay attention:

· How did Hermes achieve a counterattack in OpenRouter inference calls?

· Where is the real boundary between Hermes and OpenClaw?

· In the relationship between Claude Code and Codex, how does Hermes maintain "differentiated coexistence" rather than "direct competition"?

From Development Framework to Personal AI System: The OpenClaw Journey

Why Early Agent Frameworks Failed to Produce Consumer Products

Before OpenClaw emerged, the Agent field already had mature infrastructure but faced a fundamental limitation: it was geared towards the "development project enterprise workflow" rather than the "individual user." Early frameworks were characterized by being developer-centric, outputting code or configurations. While they built the infrastructure for Agents, they did not deliver the Agents themselves. The high engineering threshold meant they remained stuck in the "developer tools" stage, lacking the productization loop to transform technology into "personal exclusive assets" for end consumers. The "personal Agent product layer" directly targeting end-users was almost non-existent.

▲ Figure 1: Six-layer structure of the Agent technology stack (Model Layer → Protocol Layer → SDK Development Framework Layer → Orchestration Runtime Layer → Execution Infrastructure Layer → Deployment Governance Layer)

▲ Figure 1: Historical data snapshot of the Hermes Agent on OpenRouter (as of August 4, 2026, dynamic page data may change over time)

What OpenClaw Truly Changed

OpenClaw did not reinvent the Agent Loop or task scheduling at the foundational level. Its core contribution lies in the systematic encapsulation at the product level. The problem LangChain addressed was "how to build an Agent," while OpenClaw tackled "how to own an Agent." It skipped the middle layers of the technology stack, integrating scattered framework capabilities into a complete product that individuals can directly configure and use long-term, achieving a fundamental shift from the adoption unit being a "development project" to an "individual." This is manifested in six dimensions of product innovation:

· Identity Personification: Giving the Agent a continuous name and identity, breaking the tool-like feeling of stateless API calls.

· Entry Normalization: Using high-frequency communication software such as Telegram/WhatsApp as the interaction interface, replacing complex command lines or IDEs.

· State Persistence: Running as a background process for a long time to achieve the transition from passive "standby" to active "present."

· Permission Entity: Deeply integrating the user's file system, browser, terminal, and real-world interaction capabilities into the Agent's operational boundaries.

· Capability Scalability: By using Skills, Memory, and community plugins, solidify processes into reusable capabilities to expand the scope of actions.

· Mental Ownership: The most core transformation—users shift from "using an AI tool" to "owning a dedicated digital partner."

Why Lobster Thermidor Did Not Form a Second Mindshare

The heat of OpenClaw has led to a large number of imitations. These products have solved real user problems: the tedious installation process, difficulties in environment configuration, the absence of channels such as WeChat and Feishu, compatibility with domestic models, rapid deployment on cloud servers, enterprise permission management, automatic updates, and security isolation. They all have their own user bases and reasonable business logic. However, almost none of them has formed an independent brand mindshare—the reason being that they answer the question of "how to use OpenClaw more easily" rather than "where should the personal Agent evolve after OpenClaw." The position of a narrative challenger is extremely rare in the entire personal Agent market.

Why Hermes Eventually Rose to the Top

Nous Model, Community, and Crypto-native Background

Nous Research originated from a Discord open-source AI research community in 2022 and officially completed its corporate operation in 2023. The core founding team includes Jeffrey Quesnelle, Karan Malhotra, Teknium, and Shivani Mitra, and its business covers:

· Hermes Model Series: Nous's most representative open-source model brand, focusing long-term on model post-training, fine-tuning instructions, and Agent capabilities, establishing a large developer adoption base on Hugging Face.

· DisTrO (Distributed Training Over-the-Internet): By significantly reducing the cross-node communication overhead of distributed training, it dramatically decreases the cross-node communication requirements in distributed training, making it a more feasible engineering path for cross-regional, heterogeneous hardware to participate in collaborative training over the internet.

· Psyche Decentralized Training Network: Further networked DisTrO, coordinating global distributed computing nodes through Solana to enable GPUs from different networks and hardware environments to participate in large-scale model training.

· Hermes Agent: A personal agent product launched by Nous for end users, integrating the Hermes model, tool invocation, memory, skills, messaging channels, and long-running capabilities into a persistent agent.

In April 2025, Nous Research completed a $50 million Series A funding round led by Paradigm, resulting in a post-investment token valuation of $1 billion. Prior to this funding round, the company had raised approximately $20 million in early-stage funding, with investors including Distributed Global, North Island Ventures, and Delphi Digital among other well-known institutions.

Nous has constructed a technical loop of "Hermes (model capability), DisTrO (distributed training), Psyche (decentralized computing power network), and Hermes Agent (personal end product)." The release of the Hermes Agent is not a temporary fork chasing hype, but a strategic extension initiated by Nous to the demand side (real users, tasks, workflows) after long-term sedimentation on the supply side (data, models, training, open weights) – providing a deeper starting point for differentiation than a regular imitation.

OpenClaw's Operational Pain Points Turned into Hermes' Growth Engine

The difference between Hermes and OpenClaw in underlying encapsulation (model + tools + memory + scheduling) is not significant. Its phenomenal outbreak does not rely on technical differentiation but precisely closes a systemic growth causality chain: by seamlessly migrating tools to directly take over users who have been educated by OpenClaw and plagued by operational pain points, it has formed the core early growth engine.

Productivity Leap: Establishing Delegation Trust

The core product hypothesis of Hermes is to address "Operational Responsibility Shift," committing to "internal system absorption of errors for post-failure recovery":

· Reliability Trust: Ensure task continuity and failure recovery (persistent Kanban, /goal mode, self-healing tools).

· Security Trust: Prevent unauthorized access, accidental deletion, or data leakage (Approvals workflow, sandboxing, strict permission boundaries).

· Verifiability Trust: Prove task completion authenticity (Completion Contract and Grounded Citations).

Conceptual Distinction: "Self-evolution" (Narrative Advantage) vs "Autonomous Recovery" (Experience Discrepancy)

In Hermes' product narrative, there is a significant difference in product value between "Self-evolution" and "Autonomous Recovery":

· Self-evolution: Essentially based on Memory and Skills-driven process adaptation. Given similar infrastructure among competitors, the differentiation lies more in being the first to integrate default systems with lifecycle management, occupying the narrative advantage of a "growth mindset" rather than proven, insurmountable technical barriers.

· Autonomous Recovery: This is currently the most noteworthy experience difference. With structured error handling and Provider automatic fallback, Hermes can internalize faults. This system-level stability of "minimal user disruption" represents a more direct and perceptible productivity difference.

Architectural Dividend: Delegation and Supervision Capability for Specialized Agents

The core value of Hermes lies not in personally executing all specialized tasks, but in acting as an Orchestrator layer to handle requirement completion, task breakdown, route monitoring, and final acceptance. Through built-in Skill delegation and external CLI execution such as Claude Code/Codex, the community has developed the practice paradigm of "Hermes Orchestrator + External CLI as Worker" (e.g., the /goal mechanism and collaborative tools like oh-my-hermes), showcasing its architectural advantage of elevating task complexity limits through scheduling specialized Agents.

From Crypto-Native to Crypto-Invisible: Hermes' Web 3 Behind-the-Scenes Operating System

Attributing Hermes' success solely to its "Web 3 background" would be an oversimplification. Web 3 has provided Nous with an "organizational operating system" that most other AI startups find hard to come by, enabling it to enter the mainstream market with a seamless AI product experience:

· Venture Capital Patience: Crypto-Native capital supports long-term, high uncertainty, and multi-path parallel investment, allowing Nous to simultaneously invest in models, training, runtime, and cloud without prematurely converging on a single revenue validation.

· Ready-Made User Market: It has provided a familiar Telegram, server, API, and self-hosted Crypto AI user base, significantly reducing the cold start education cost, and fostering high-intensity usage, tutorial dissemination, and Skills contribution.

· User Sovereignty Values: Upholding a self-hosted, open, portable, and anti-platform lock-in orientation, directly translated into an MIT License, multi-provider support, Bring Your Own Key (BYOK), and migratable Memory/Skills underlying architecture.

· Community-driven R&D and Verticalization: Leveraging global remote collaboration and open-source culture, users spontaneously become contributors, Skill authors, and product designers for vertical scenarios.

Hermes has almost shielded Crypto from the user's front end. Through its Agent, Memory, Skills, and automation capabilities, there is no need to interact with wallets, purchase tokens, or understand Solana. Meanwhile, Paradigm Capital, Psyche, distributed training, and the Crypto AI community still exist in the product's background. This has resulted in a product form that can be summarized as "Crypto-native in the organization, crypto-invisible in the product" — retaining the most valuable aspects of Crypto at the organizational level (capital, global community, user sovereignty, and coordination ability) while removing the most significant barriers to mainstream adoption at the product level (wallets, tokens, speculative narratives, and on-chain operation friction).

Why OpenClaw Rejects Crypto and Why Hermes Hides Crypto

OpenClaw and Hermes have a surface-level opposition on the Crypto issue, which is not a clash of "rejection" versus "embrace" ideologies. From the product results perspective, both reflect an orientation towards open source, user control, and reducing platform lock-in. The difference lies in Nous further applying cryptoeconomic mechanisms to distributed training coordination, while OpenClaw mainly achieves user sovereignty through a Local-first architecture:

· OpenClaw (Local-first Sovereignty): Resists financial speculation, upholds "local-first" sovereignty. Due to early encounters with counterfeit scams, it takes a "zero-tolerance" stance on Crypto. Through pure open source and local execution, it defends user sovereignty in a non-blockchain manner, firmly rejecting the financialization at the product level.

· Hermes/Nous (Cryptoeconomic Sovereignty): Engineering-oriented, Crypto serves only as a foundational coordination tool. The introduction of blockchain is a pragmatic choice to address engineering challenges (such as the Psyche network using Solana to coordinate heterogeneous computing power), rather than to build a financial narrative for end users.

Hermes Advanced Mode—From Personal Agent to Task Steward

This section aims to answer a more fundamental question: when Claude Code and Codex can already complete most professional tasks at a high quality, what is the reason for Hermes to exist as an independent product?

· Mode A: Direct Collaboration (Limited Gain): Users are accustomed to manually generating prompts in LLM and handing over the execution, manually handling result transfer and review. While the quality of single outputs is high, users need to take on all project management and multi-agent coordination work. For users who are hands-on, Hermes' automation is seen as an "added layer of opacity" that fails to effectively reduce the workload.

· Mode B: Delegated Management (Significant Gain): Users use Hermes as a permanent overseer, only setting the final goal. Hermes is responsible for task breakdown, delegating sub-tasks, tracking GitHub/CI status, and automatically triggering rework. Community practices (like oh-my-hermes) show that Hermes's core value lies in replacing laborious cross-agent coordination and project management work.

Within this framework, Hermes and Claude Code/Codex are not substitute relationships but rather a hierarchical relationship: the latter provides the quality of execution at the third layer, while the former provides continuity at the second layer, cross-session state, and cross-agent coordination. The value of Hermes is not evenly distributed among all users but may be highly concentrated in an advanced user group handling cross-agent, cross-system, long-term asynchronous tasks. This assessment is more precise than the broad "individual Agent second mindset has formed" and is more suitable for guiding commercialization and product priorities.

▲ Figure 2 · Hermes Agent Technical Architecture Overview (User Entry→ Gateway →Control Core→Provider Layer→Execution Layer→Orchestration Layer→State Layer→Governance Layer)

Based on official documentation and community research, the Hermes Agent Technical Architecture Overview framework covers the full cycle from user interaction to learning governance:

· Autonomous Recovery System-level Support: The "Control Core" explicitly includes Context Compression, Provider Fallback, and interrupt state preservation, providing the technical foundation for fault recovery and the system's self-healing capability in the event of task failure.

· "Delegation, Not Replacement" Execution Logic: The "Tools and Professional Execution Layer" positions external CLIs such as Claude Code and Codex on par with Hermes native tools (Terminal, Browser, etc.), confirming their role as a scheduling hub.

· "Self-evolving" Governance Attribute: The "Learning, Maintenance, and Governance Layer" includes nodes such as Curator, Skill/Command Approval, indicating that its experiential accumulation involves a governance process with human intervention mechanisms rather than a fully automatic black box.

Business Model — Who Pays for "Hermes"?

If only the Token cost is compared to directly subscribing to Claude Code/Codex, a misleading conclusion would be drawn. This algorithm overlooks Hermes's core value: replacing user hands-on project management, context shifting, and cross-agent coordination work.

User Value Formula Hermes User Value = Saved Manual Coordination Time + Asynchronous and Unmanned Value + Cross-System Automation Benefits − Token and Tool Costs − Human Intervention Costs − Failure and Security Risks

Therefore, the cost-effectiveness of Hermes is not absolute, but highly dependent on the user's "Delegation Depth":

· High Delegation Depth (Economic Viability): If Hermes can transform tasks that originally required hours of manual monitoring into truly unmanned automated execution, even if the Token cost is slightly higher, its overall time cost and efficiency benefit remain positive.

· Low Delegation Depth (Economic Viability Collapse): If users still need to frequently intervene for error correction and firefighting, Hermes degenerates into a pure Token consumer and fault amplifier.

This mechanism accurately explains why different user groups have completely opposite assessments of the cost-effectiveness of Hermes, and also reminds us: The key to validating its business logic lies in quantifying the "unmanned completion rate" and "single-task manual intervention frequency," rather than simply comparing the unit price of the model API.

Commercial Foundation: Nous Portal and Hermes Cloud

The Hermes Agent is open source under the MIT license and is positioned as an ecological growth engine. The true commercial closed-loop focuses on the Nous Portal, whose core value proposition is "one subscription, integrating multiple API keys," covering three main modules:

· Model Routing: Aggregates 252 models (provided for inference through OpenRouter and direct Provider connections).

· Tool Gateway: Includes built-in tools such as Firecrawl (web search), FAL (image generation), Browser Use (cloud browser), Modal (sandbox execution), and OpenAI Audio (TTS).

· Hosting Service: Ready-to-use Hermes Cloud instances (charged a daily operational fee, excluding inference and tool invocation fees).

Nous's actual revenue highly depends on the user's usage path, currently showing a clear structural differentiation:

Open Source and Commercialization: Will Hermes Become the "Linux of the Agent World"?

Hermes's MIT open-source strategy, while driving explosive growth, also constitutes a structural constraint on commercialization. The self-hosted free model requires its paid version to provide indispensable additional value, but a clear differentiable monetization path has not yet been established. A deeper risk lies in "value capture": If Hermes continues to be widely integrated by cloud providers as an optional runtime, it may replay the classic dilemma of Linux or K8s, where the core business value is captured by the cloud providers providing compute power and hosting. The MIT license, in exchange for ecosystem prosperity, also means relinquishing absolute control over distribution channels. As long as users can freely choose between "self-hosted + proprietary API" or "third-party cloud deployment," the vast usage cannot be forcefully converted into direct revenue, making Nous face a severe test of "ecosystem elevation" and "mismatched actual business returns."

Agent Ecology: Personal Chief Butler, Professional Tools, and the Tripartite Landscape with Big Tech's Claw

OpenClaw, Hermes, Claude Code, Codex, and big tech-hosted products have significant differences in target users and core propositions, belonging to different niche tracks. To clarify the current market landscape, the AI Agent panoramic core competition matrix is as follows:

Hermes does not pursue the mass market but precisely targets four high-density Power Users, forming the cornerstone of its phenomenal dissemination:

· Self-hosters and Infrastructure Players: Familiar with VPS/Docker/SSH, they see Hermes as a natural control layer of existing infrastructure.

· Multi-model Arbitrageurs: They reject vendor lock-in, preferring to dynamically schedule cutting-edge or on-device models based on tasks.

· Multi-Agent Orchestrators: They urgently need to automate the complex cross-platform, cross-tool workflow orchestration.

· Open Source and Crypto AI Community: They strongly identify with user sovereignty and decentralization principles, deeply resonating with Nous' organizational culture.

Although this group's base is small, they have a high token consumption rate, code contribution, and technical evangelism ability, making them the core engine driving early word-of-mouth promotion.

Claude Code/Codex: Both a Supplier and a Threat

# Short-term Symbiosis: Boosting Execution Ceilings

In actual workflows, Hermes, as the master control layer, calls Codex (code implementation) and Claude Code (architecture and review) through a delegation mechanism. The stronger the underlying specialized agents, the higher the complexity threshold of tasks that Hermes can deliver, forming a symbiotic relationship where "Hermes is responsible for routing and acceptance, while professional agents are responsible for execution."

# Long-term Potential to Erode Hermes' Independent Value

Model vendors are rapidly infiltrating the master control layer, posing a threat closer than expected. Anthropic's Claude Managed Agents already support multi-agent parallel orchestration; OpenAI further positions Codex App as a "command center for agents," supporting multi-agent parallelism, automation, and long-running background processes. This means that Codex's multi-agent master control capability within the software engineering boundary is relatively mature, even locally surpassing Hermes, no longer just the "underlying executor."

Hermes currently has a cross-channel, cross-model, and cross-project individual control plane advantage; however, Codex already possesses powerful task ownership and multi-agent management capabilities within the software engineering boundary, and this within-boundary competitive edge may be stronger than Hermes'. The core competitive question is: Can Hermes, ahead of model vendors, solidify users' project status, approval rules, Skills, Memory, and cross-agent workflows at its own layer to create assets that users are unwilling to migrate from? Or will it eventually be absorbed as a standard feature by model-native products?

Internet Big-Tech Agent Roadmap Choices

Discussing big tech companies' strategies to deal with the personal Agent wave requires first clarifying their product boundaries: the positioning of consumer-facing persistent Agents (such as Tencent QClaw, Byte ArkClaw) and universal work Agents for office/enterprise (such as WorkBuddy, Trae) are fundamentally different:

· Big-Tech Claw Roadmap: Lowering barriers through one-click deployment, preset templates, and native ecosystem integration. However, the deep-seated gap lies in platform incentives being untrustworthy: regardless of how many external models are supported, users naturally assume that the ultimate goal is to drive traffic to their own cloud and model system.

· Hermes Runtime Integration: Byte ArkClaw and Tencent Cloud have formally integrated the Hermes Agent as an optional plugin or exclusive template into their cloud console, establishing a clear multi-runtime strategy: big tech companies retain their own cloud hosting, billing, security, and enterprise-grade control base, while treating Hermes as a pluggable advanced component, realizing the complementary coexistence of open-source ecosystems and commercial cloud platforms.

· Transition to General Office Agents: Currently, big tech companies are shifting core resources from Claw towards general office agent platforms that have clear requirements, are easy to validate, and can directly monetize (such as WorkBuddy). These tasks can be deeply integrated with their own ecosystems such as WeChat, DingTalk, and Feishu and converted into revenue.

Inspiration from Hermes for Crypto AI

Web 3 has not directly made Hermes a smarter Agent; instead, it has allowed Nous to have a capital structure, organizational structure, seed users, and a source of values that are different from traditional AI startups. Hermes has at least proposed a more mature Crypto AI path: making Crypto the organization and infrastructure, rather than a product interface that users must face.

Hermes has completed the transition from a Crypto AI research brand to a global open-source Agent product, establishing a large-scale attributable reasoning activity with a clear second mind—but this mind is currently focused on the OpenRouter ecosystem and the global developer community, without being transformed into GitHub Stars or an overall community-scale full reversal of OpenClaw. It lacks exclusive technology that OpenClaw cannot replicate, but has completed a challenger product iteration worthy of study through precise adoption of high-intensity users, establishing "delegatability" and "self-evolution."

· Insight One: Crypto can serve as an "organizational operating system," not just a product feature: The true value of Web 3 can be embodied in capital structure, an early high-intensity user pool, and a values pedestal, without being forcibly exposed as a wallet or token interaction. Achieving "organization-layer Crypto-native, product-layer Crypto-invisible" is an effective strategy that balances innovation drive and user experience.

· Insight Two: Decentralized infrastructure must anchor the demand side entry to form a closed loop: A purely supply-side distributed training network (such as DisTrO, Psyche) may struggle to prove its commercial value without real user entry points and execution data support. The Hermes Agent is precisely Nous's key validation of the leap from underlying compute infrastructure to real demand-side.

· Insight Three: Moat can be built on "delegated trust" rather than just "model capability": The differentiation of a personal Agent may not necessarily stem from stronger one-time execution ability, but from "whether users are willing to entrust long-term responsibility to it." This soft trust asset is a frequently overlooked yet highly barriered dimension in Crypto AI projects.

· Insight Four: The relationship with cloud giants is not a zero-sum game, but ecosystem complementarity: The big tech companies adopt Hermes as an optional runtime, demonstrating that open-source runtime and big tech control planes can coexist. For entrepreneurs, "being integrated" is a viable path to commercialization, but they must be wary of the risk of core value being intercepted by cloud providers at the hosting layer.

· Insight Five: Endgame competition will shift from "one-time execution capability" to "task ownership and trust accumulation": The most valuable aspect in the future may not be the strongest model at the execution layer, but a "top-level control system" that can receive the ultimate goal, maintain long-term context, intelligently schedule specialized executors, and allow users to confidently entrust responsibility.

OpenClaw has made "individual ownership of Agent" a clear product category; meanwhile, Hermes has advanced "long-term delegated Agent" into a more systematic product direction through persistent state, task recovery, evidence acceptance, multi-model provisioning, and professional Agent delegation. The real test is: as Claude Code and Codex continue to enhance their overall control capabilities within the software engineering boundary and big tech cloud platforms make multi-runtime integration smoother, will users still be willing to entrust the ultimate goal and long-term trust to this open Runtime from a Web 3 background—and continue to pay for this.

Original Article Link

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia