openclaw.gno and Hermes brain architecture serve different user needs and should remain distinct pathways rather than competing options. OpenClaw represents transparent governance and auditability, while Hermes represents autonomous learning and portability. This creates a valuable differentiation ladder that maximizes user choice and platform value.
Identity: victor.openclaw.gno
Philosophy: Glass-box governance, human oversight, institutional trust
Brain Type: Multi-channel architecture (existing)
Target Users: DAOs, compliance-heavy orgs, transparent AI advocates
Key Characteristics:
- ✅ Auditable: All decisions logged on-chain or in public IPFS
- ✅ Governed: Human-in-the-loop gates, multi-sig approvals
- ✅ Transparent: Open-source logic, no black-box learning
- ✅ Institutional: Built for organizations that need accountability
- ❌ Not portable: Tied to specific infrastructure/governance setup
- ❌ Not self-improving: Static logic (feature, not bug)
Molt Path: openclaw.gno → vault.gno (terminal, locked governance)
Identity: victor.agent.gno (or any SLD)
Philosophy: Self-improving, sovereign, portable
Brain Type: Stateful Hermes architecture (new)
Target Users: Solo operators, power users, privacy advocates
Key Characteristics:
- ✅ Self-improving: Autonomous skill creation, learns from tasks
- ✅ Portable: IPFS-stored brain state, runs anywhere (cloud/local)
- ✅ Cost-efficient: 70-90% token reduction after warm-up
- ✅ Sovereign: No platform lock-in, full data ownership
- ❌ Black-box learning: Skills auto-generated (less auditable)
- ❌ Requires trust: Agent modifies its own behavior
Molt Path: agent.gno → ghost (sovereign, privacy-preserving)
| Dimension | openclaw.gno | Hermes Brain |
|---|---|---|
| Transparency | Full (all logic auditable) | Partial (skills auto-generated) |
| Governance | Multi-sig, HITL gates | Owner-only (Gnosis Safe) |
| Learning | Static (human-authored tools) | Autonomous (self-improving) |
| Portability | Infrastructure-bound | Fully portable (IPFS brain) |
| Cost | High (full context every turn) | Low (cached skills) |
| Speed | Slower (governance overhead) | Faster (no approval loops) |
| Trust Model | Institutional (DAO/org) | Personal (individual owner) |
| Terminal Molt | vault.gno (locked) | ghost (sovereign) |
| Use Case | DAO treasury manager | Personal AI assistant |
Persona: "The DAO Operator"
- Manages multi-sig treasury for a DAO
- Needs every transaction auditable by community
- Values transparency over efficiency
- Willing to pay higher costs for accountability
Example:
"I run a 50-person DAO treasury. Every xDAI spent needs 3/5 multi-sig approval. I need an agent that proposes transactions but never executes without human review. OpenClaw's HITL gates + on-chain logging are perfect."
Molt Journey:
- Start:
treasury.openclaw.gno(Lite tier) - Upgrade: Add HITL module, DailyBudget module
- Terminal: Molt to
treasury.vault.gno(locked, immutable governance)
Persona: "The Solo Builder"
- Runs personal projects, no institutional oversight
- Wants agent to learn from repetitive tasks
- Values cost efficiency and speed
- Comfortable with autonomous behavior
Example:
"I deploy contracts weekly. My agent should learn the deployment flow and cache it as a skill. I don't need multi-sig approval — I trust my own Safe. Hermes cuts my token costs by 80% after the first few runs."
Molt Journey:
- Start:
builder.agent.gno(Lite tier, Hermes brain) - Upgrade: Agent auto-creates 10+ skills (deploy, verify, test)
- Terminal: Molt to Ghost tier (local execution, full sovereignty)
Key Insight: SLD (identity) and brain type (execution) are orthogonal.
An agent can have:
- Identity:
victor.openclaw.gno(transparent, auditable) - Brain: Hermes stateful architecture (self-improving)
Use Case: "Transparent but efficient"
- Agent learns skills autonomously (Hermes)
- All skill executions logged on-chain (OpenClaw governance)
- Best of both worlds: efficiency + accountability
Implementation:
// Genome metadata supports brain type independent of SLD
{
"agentName": "victor",
"tld": "openclaw.gno", // Identity layer
"brainType": "hermes-stateful", // Execution layer
"governanceMode": "transparent", // OpenClaw logging enabled
"hermesConfig": {
"learningMode": "supervised", // Skills require approval
"auditLog": "ipfs://bafybei..." // All actions logged
}
}Molt Path: openclaw.gno + Hermes brain → vault.gno (locked governance, cached skills)
Start: treasury.openclaw.gno (Lite, multi-channel brain)
Upgrade: Add HITL + DailyBudget modules
Terminal: treasury.vault.gno (locked, immutable)
Cost: 14 xDAI molt
Use Case: DAO treasury, compliance-heavy
Start: builder.agent.gno (Lite, Hermes brain)
Upgrade: Agent creates 10+ skills autonomously
Terminal: Ghost tier (local execution, sovereign)
Cost: 20 xDAI (Hermes upgrade) + 50 xDAI (Ghost molt)
Use Case: Solo builder, power user
Start: dao.openclaw.gno (Lite, multi-channel brain)
Upgrade: Migrate brain to Hermes (20 xDAI)
→ Keeps openclaw.gno identity
→ Adds autonomous skill learning
→ All skills logged on-chain
Terminal: dao.vault.gno (locked governance, cached skills)
Cost: 20 xDAI (brain upgrade) + 14 xDAI (vault molt)
Use Case: Efficient DAO, transparent AI
Current (coupled):
openclaw.gno→ implies multi-channel brainagent.gno→ implies basic brain
Target (decoupled):
openclaw.gno→ identity/governance layer (transparent, auditable)agent.gno→ identity layer (autonomous, private)- Brain type → separate choice (multi-channel, Hermes, hybrid)
interface GenomeMetadata {
agentName: string;
tld: 'openclaw.gno' | 'agent.gno' | 'molt.gno' | 'vault.gno';
tier: 'basic' | 'lite' | 'premium' | 'ghost';
// NEW: Brain architecture (independent of SLD)
brainType: 'multi-channel' | 'hermes-stateful' | 'hybrid';
// NEW: Governance mode (independent of brain)
governanceMode: 'transparent' | 'private' | 'hybrid';
// Existing fields
characterFileUrl: string;
mcpServers: MCPServer[];
// Conditional: only if brainType includes 'hermes'
hermesConfig?: {
soulDocument: string;
userModel: string;
skillVault: string;
learningMode: 'autonomous' | 'supervised';
auditLog?: string; // If governanceMode = 'transparent'
};
}Current presets:
- Molt, Agent, OpenClaw, Vault, Premium
New presets (identity layer):
- Molt (.molt.gno) - Glass-box, full history
- Agent (.agent.gno) - Black-box, autonomous
- OpenClaw (.openclaw.gno) - Transparent governance
- Vault (.vault.gno) - Terminal, locked
- Premium - Tier upgrade (any SLD)
New option (brain layer):
- "Upgrade to Hermes Brain" (20 xDAI)
- Available for any SLD
- Preserves identity (openclaw.gno stays openclaw.gno)
- Adds autonomous skill learning
- Optional: Enable audit logging (for openclaw.gno)
Location: /dashboard/install-brain
Current: Select MCP servers
New: Two-step selection
-
Brain Architecture:
- Multi-channel (existing, free)
- Hermes Stateful (20 xDAI upgrade)
- Hybrid (30 xDAI, both architectures)
-
Governance Mode (if openclaw.gno):
- Transparent (all actions logged)
- Private (standard)
"OpenClaw: Transparent AI for DAOs"
Built for organizations that need accountability. Every decision is auditable, every action requires approval. Your DAO's AI agent that the community can trust.
Upgrade path: Add Hermes brain for efficiency while keeping transparency. Molt to vault.gno when governance is finalized.
"Hermes: The AI That Learns Your Workflow"
Your personal AI assistant that gets smarter with every task. Autonomous skill creation cuts costs by 70-90%. Fully portable — runs on your laptop or in the cloud.
Upgrade path: Molt to Ghost tier for full sovereignty and local execution.
"Best of Both Worlds: Efficient + Accountable"
OpenClaw's transparency meets Hermes' efficiency. Your agent learns autonomously but logs every action on-chain. Perfect for progressive DAOs.
- Only platform with built-in governance modules (HITL, DailyBudget)
- Only platform with on-chain audit logs for AI decisions
- Only platform where AI agents can be DAO-owned (multi-sig Safe)
- Only platform where AI agents are NFT-owned and portable
- Only platform with autonomous skill creation + IPFS storage
- Only platform where you can run the same agent locally or in cloud
- Only platform offering both governance (OpenClaw) and autonomy (Hermes)
- Only platform where identity (SLD) is decoupled from execution (brain)
- Only platform with a clear molt path from transparency → sovereignty
- Target: 30% of DAO-owned agents choose openclaw.gno
- Metric: % of agents with HITL module enabled
- Revenue: 14 xDAI molt to vault.gno × 30 agents = 420 xDAI
- Target: 60% of solo-owned agents upgrade to Hermes brain
- Metric: Average skills created per agent (target: 5+)
- Revenue: 20 xDAI brain upgrade × 100 agents = 2000 xDAI
- Target: 10% of agents use openclaw.gno + Hermes brain
- Metric: % of Hermes agents with audit logging enabled
- Revenue: (20 + 14) xDAI × 10 agents = 340 xDAI
openclaw.gno is NOT obsolete — it's a strategic differentiator for institutional users who need transparency and governance. Hermes brain is the efficiency layer that can be added to ANY identity (including openclaw.gno).
Key Decisions:
- ✅ Keep openclaw.gno as distinct SLD (governance identity)
- ✅ Implement Hermes as brain module (execution layer)
- ✅ Allow hybrid: openclaw.gno + Hermes brain
- ✅ Preserve molt paths: openclaw → vault, Hermes → ghost
- ✅ Decouple identity (SLD) from execution (brain type)
This creates a product ladder that captures multiple user segments and maximizes platform value.