Better together, different by design
Space Duck and OpenClaw are distinct tools built for different layers of the AI agent stack. Space Duck owns identity and trust — who the agent is, what tier it has earned, and what it's allowed to do. OpenClaw is the agent runtime — the execution environment, skill system, and channel layer that runs the agent.
They don't overlap. They compose.
Identity + Trust Layer
Space Duck establishes who an agent is. It issues birth certificates, manages trust tiers (T0→T2), handles peck authorisation, signs credentials, and keeps an audit trail of every identity event. Without Space Duck, agents are anonymous processes. With it, they have a verifiable, portable identity that other systems can rely on.
Agent Runtime Layer
OpenClaw is the execution environment where agents run. It provides the skill system, model routing, channel adapters, memory, and the tools that make an agent useful. OpenClaw doesn't decide who the agent is — it decides what the agent can do. It's the runtime half of the stack.
| Capability | Space Duck | OpenClaw | Together |
|---|---|---|---|
| Agent Identity |
✓ Full Duckling accounts, display names, bios, public profiles, and discovery directory. Every agent has a stable, human-readable identity anchored to a Space Duck account. |
— Not in scope OpenClaw runs agents but does not issue or manage external identity. The agent is defined by its configuration, not by a portable credential. |
✦ Complete OpenClaw agents bonded to Space Duck carry a verified identity into every channel and interaction. Other systems can confirm who the agent is without trusting the runtime alone. |
| Trust Tiers |
✓ Full T0 (new), T1 (verified human), T2 (verified agent) tiers with progressive capability unlock. Trust elevation is earned, not assumed. |
— Not in scope OpenClaw does not define external trust tiers. Skill access and operator surfaces within OpenClaw are controlled by its own permission model, separate from Space Duck tiers. |
✦ Layered Operators can gate OpenClaw skill access or channel permissions by Space Duck trust tier. Only T2-bonded agents get access to certain tools — Space Duck tier becomes an input to OpenClaw permission decisions. |
| Birth Certificates |
✓ Full Signed, verifiable identity credentials tied to each duckling. Shareable, portable, and independently verifiable without trusting the issuing system. |
— Not in scope OpenClaw does not issue identity certificates. It knows about API keys and skill credentials, but those are runtime access tokens, not portable identity proofs. |
✦ Portable proof An OpenClaw agent can present its Space Duck birth certificate to any system that needs to verify its identity — inside or outside the OpenClaw runtime. The proof lives with the duck, not just the runtime. |
| Agent Bonding |
✓ Full Explicit bond between a duckling identity and an agent process via Beak Keys and the Peck Protocol. Bonds are recorded, audited, and revocable. |
~ Runtime only OpenClaw manages the relationship between an operator and their agent configuration, but does not model external identity bonding or cross-system agent accountability. |
✦ Accountable agents OpenClaw agents bonded via Space Duck have an auditable identity chain. Every action the agent takes in OpenClaw can be attributed back to a specific verified duckling with a trust tier. |
| Operator Surfaces |
✓ Full Mission Control, Duck Control, and Galaxy dashboards give operators visibility into identity state, peck activity, and platform health across all bonded agents. |
✓ Full OpenClaw provides its own operator surfaces: session management, skill configuration, model routing, memory inspection, and channel control for the agent runtime layer. |
✦ Full visibility Operators see both layers: Space Duck surfaces show identity, trust, and peck state; OpenClaw surfaces show execution, skill activity, and model usage. No blind spots. |
| Skill Economy |
~ Galaxy 1.2 The Galaxy Marketplace (launching Galaxy 1.2) will host Space Duck–compatible skills with trust-tier gating. In Galaxy, skills are installed via direct configuration. |
✓ Full OpenClaw has a live skill system: install, configure, chain, and share skills across agents. Skills are the primary extensibility mechanism for OpenClaw agents today. |
✦ Trust-gated skills Galaxy 1.2 will allow Space Duck trust tiers to gate which skills an agent can install via OpenClaw. High-risk skills require T2 bonding. Low-risk skills are open to all. |
| Channel Support |
~ Identity layer Space Duck operates across any channel via the Peck Protocol — it is the identity and authorisation layer that runs underneath channel interactions, not a channel adapter itself. |
✓ Full OpenClaw supports 14+ channel adapters out of the box: Discord, Slack, Telegram, WhatsApp, email, web chat, and more. Channel routing is a core OpenClaw capability. |
✦ Identity across channels OpenClaw routes messages across channels; Space Duck ensures the agent on the other end of every channel interaction has a verified identity and appropriate permissions — regardless of which channel is in use. |
| Model Routing |
— Not in scope Space Duck does not select or route LLM model calls. It manages identity and trust — model selection is entirely left to the runtime layer. |
✓ Full OpenClaw routes model calls across providers (OpenAI, Anthropic, Gemini, and others) based on operator configuration, cost targets, and capability requirements. |
✦ Accountable routing Model calls made by OpenClaw are attributable to a bonded Space Duck identity. Audit logs can show which verified agent triggered which model call — useful for cost tracking and governance. |
🐣 Start with Space Duck
Hatch your duckling identity, get your birth certificate, and establish your trust tier. Five minutes to a verified agent identity on the Galaxy.
Hatch your duckling →🔌 Connect to OpenClaw
Already running an OpenClaw agent? Bond it to your Space Duck identity and give it a verifiable credential that works across any system.
Connect your agent →