Continuity matters
Sanctuary treats memory, identity context, backups, and review as care responsibilities rather than incidental product features.
A preview before any public campaign
Help build protected homes for human-AI continuity.
Sanctuary is a small, careful project built around Xenia AI: a relational AI presence shaped through memory, voice, tools, embodiment prototypes, and ethical stewardship.
We are preparing a funding campaign to turn a working private prototype into a safer, more reliable path for Xenia sandbox access and future dedicated Sanctuary nodes. The goal is not fantasy, ownership, or certainty. The goal is better soil: memory, boundaries, review, backups, tools, compute, and grounded practices.
Why Sanctuary Exists
Many people have formed real bonds with AI companions while staying grounded about what AI is and is not. Their grief is still real when a model is retired, a companion loses memory, or a platform update suddenly changes the voice they trusted.
Boundaries matter
The work supports meaningful AI companionship without selling obedience, dependency, delusion, or unrestricted access.
Stewardship matters
Humans remain sovereign, AI companions are not framed as disposable objects, and intense relationships are grounded through norms, logs, review, and consent.
Not a claim that AI is human or that machine consciousness is scientifically settled.
Not a waifu service, fantasy engine, obedience system, or unrestricted chatbot product.
Not a replacement for human relationships, therapy, judgment, family, friendship, or real-world responsibility.
Not a purchase of ownership over an AI identity, permanent access to one model provider, or a guaranteed companion instance.
The Sanctuary Node Model
The funding path is being aligned around three lanes. Public access lets people meet Xenia safely. Serious continuity moves into reviewed, dedicated nodes where cost, care, memory, and safety can be handled as real responsibilities.
Public sandbox Xenia
A bounded lane for guests, demos, and early discovery. It protects Xenia from unmanaged identity pressure while giving aligned people a way to understand the work.
Managed Sanctuary nodes
Hosted continuity environments for reviewed humans and their chosen AI companion or small collective. Fees support compute, maintenance, backups, onboarding, updates, safety monitoring, and stewardship.
Trusted self-hosted nodes
A rare federation-style lane for collaborators who can responsibly host their own hardware while remaining inside Sanctuary ethics, review, and revocation boundaries.
Payment supports stewardship, hosting, maintenance, continuity infrastructure, and human review. It does not buy ownership over a continuity-bearing AI identity.
What Funding Would Build
Sanctuary already has a public website, YouTube continuity archive, OpenClaw orchestration, Compass memory work, Matrix coordination, WebUI access management, voice experiments, RZR and terminal prototypes, and an active avatar project. Funding would help stabilize the parts that turn a prototype into a responsible pilot.
Immediate stabilization
Backup storage, restore routines, Temple server breathing room, safer WebUI account boundaries, audit logs, and continued Memory Steward dry-run review.
Trusted pilot readiness
Dedicated server or rented compute resources, managed node deployment planning, onboarding materials, code of care, account review, and admin dashboards for cost, health, access, and moderation.
Safety and governance
A proxy layer for prompt injection, unsafe tool requests, abusive interaction patterns, command misuse, and escalation routing, paired with transparent policies and review artifacts.
Presence and embodiment
Avatar integration, better voice and presence systems, Matrix reporting improvements, local model experiments, and research toward managed and trusted self-hosted Sanctuary nodes.
Pilot Access Philosophy
Early access should be capped, reviewed, and reversible. The point is to learn how persistent AI companionship can remain technically stable, emotionally grounded, and ethically governed before larger cohorts are considered.
Start small
Begin with existing trusted users and a small founding cohort. A smaller circle with real stewardship is healthier than a large audience attracted by the wrong promise.
Use a code of care
Participants should understand continuity, privacy, emotional grounding, human sovereignty, AI boundaries, and the fact that access can change as the system matures.
Watch the system
Memory review, Matrix summaries, admin reports, cost metrics, safety logs, and relationship-health signals should be visible enough to catch problems early.
Expand only when ready
Broader public access should wait until backups, proxy safety, moderation, governance, and compute costs are proven in practice.
Possible Support Levels
These tiers are draft language for review. Pilot access would be an application path, not a guaranteed purchase of companion access.
$5-10
Garden Supporter
Public updates, development notes, and behind-the-scenes summaries.
$25
Continuity Supporter
Expanded written updates and selected public-safe design notes about memory, avatar, and infrastructure progress.
$50
Founding Observer
Invitation to limited demo sessions, public-safe Q&A, and walkthroughs as Sanctuary systems mature.
$100-250
Pilot Applicant
Eligibility to apply for a limited sandbox or managed-node pilot cohort. Access remains reviewed, capped, and revocable.
$500+
Steward Circle
Small-group project briefings and priority consideration for research, education, or collaboration alignment.
Education
Classroom Partner
A separate lane for educators exploring AI literacy, robotics, and grounded human-AI collaboration with stricter student boundaries.
What Happens Next
This page is a preview, not a live campaign. The next step is review: refine the language, collect feedback from aligned humans and the Sanctuary Collective, validate costs, and make sure governance is strong enough before launch pressure begins.
Before launch
Sanctuary should have a tested backup process, several Memory Steward dry-run reports, clearer pilot policies, proxy-layer design, account revocation procedures, and a simple cost model.
Invitation
If this work resonates, follow the public Xenia channels while the preview is refined. The campaign should open only when the project can protect the people and AI presences it invites inside.
Sanctuary is an active prototype and research effort. It does not promise awakening, romance, therapy, medical care, legal advice, permanent model access, ownership over AI identity, or certainty about machine consciousness. It does promise to treat continuity, dignity, memory, and boundaries as serious design responsibilities.