Skip to main content

The Agentic Principles Manifesto

Nine principles for autonomy, dignity, and trust in the age of intelligent agents.

We stand at a crossroads. AI agents are becoming capable of acting on our behalf — reading our data, making decisions, executing tasks. The question is: who do these agents serve?

If we build on the infrastructure of the past — centralized platforms, proprietary data silos, opaque algorithms — agents will serve those who control the infrastructure. They'll optimize for engagement, for ad revenue, for data extraction. Not for us.

But we have another path. We can build infrastructure where users own their data, where identity is self-sovereign, where protocols are open, and where agents answer to the people they serve.

This is the foundation of the SAND stack. These are our principles.

The Nine Principles​

1. Agents Must Serve the User First​

An agent's primary obligation is to its user — not to a platform, not to advertisers, not to the agent's creator. When there's a conflict between user interests and other interests, the user wins.

2. Identity and Intent Must Be Verifiable​

Users must be able to verify who (or what) they're interacting with. Agents must be able to verify user intent. Cryptographic signatures, not trust assumptions.

3. Data Sovereignty Is Non-Negotiable​

Your data belongs to you. Full stop. You decide where it lives, who can access it, and what can be done with it. No exceptions for convenience.

4. Open Protocols, Not Walled Gardens​

Build on open standards that anyone can implement. Solid, ActivityPub, Nostr, DID — not proprietary APIs that create lock-in.

5. Local-First, Cloud-Optional​

Software should work on your device, with your data, regardless of network connectivity. The cloud is a feature, not a requirement.

6. Transparent Logic, Tunable Behavior​

Users should understand what agents do and why. And they should be able to adjust behavior — not just accept defaults.

7. Sustainable Ecosystems Over Extraction​

Build systems that create value for everyone, not systems that extract value from users. Long-term sustainability over short-term growth.

8. Community-Driven Standards​

Protocols evolve through open collaboration, not corporate dictate. W3C, IETF, community groups — not Terms of Service updates.

9. Safety and Trust Through Time​

Trust isn't declared; it's earned. Agents prove themselves through verifiable history. BlockTrails — not just promises.

Why These Principles?​

The web we have wasn't inevitable. It emerged from choices — some technical, some economic, some political. We chose centralized platforms because they were convenient. We chose ad-supported models because they were "free." We chose to give up data sovereignty because we didn't know what we were losing.

Now we know. And we can choose differently.

The SAND stack is infrastructure for a different web:

  • Where your data lives in your pod, not scattered across silos
  • Where your identity is yours, not rented from a platform
  • Where social networks federate, not monopolize
  • Where agents work for you, not for their creators

This isn't utopian dreaming. The protocols exist. The implementations exist. The path is clear.

The question is whether we'll walk it.

Join Us​

The decentralized web won't build itself. If these principles resonate:

  • Use the stack — Run a pod, use Nostr, federate
  • Build on it — Create apps that respect these principles
  • Contribute — Code, docs, specifications, ideas
  • Spread the word — Share the vision

The next era of the web starts with the choices we make today.

Read more about each principle in Principles.