Personal Continuity Layer . by RavelMinds
RavelMinds

Portable continuity for the systems you choose to keep.

RavelMinds is a boutique mind-layer entity. Its Personal Continuity Layer keeps stable, bounded behavioral context portable across tools, models, and time without claiming identity transfer, consciousness, or emotional equivalence.

Behavior, not identity . Continuity, not personhood . Portable layer, not platform lock-in

Read the layer -> See current status
Front door
01 . Layer
Start with the continuity packet.
See the bounded object RavelMinds intends to carry: preferences, rituals, retrieval posture, drift controls, and explicit exclusions.
Open layer
02 . Doctrine
Read what RavelMinds will not claim.
The product posture remains load-bearing only if companion-style use stays bounded by no-identity-transfer, no-dependency, and portability rules.
Read doctrine
03 . Status
See where the build actually stands.
The route exists now. The runtime does not. Study, packet design, and consent quality still lead the next gate.
View status
Continuity layer

RavelMinds carries continuity as a bounded layer, not as a synthetic self.

The layer is intentionally small. It should carry enough structure to preserve stable behavior across sessions and systems while staying inspectable, portable, and easy to refuse when source quality or consent quality is weak.

Packet
Continuity packet
A bounded packet holds durable preferences, approved memories, interaction shape, and explicit exclusions. The packet moves. The person does not get collapsed into the model.
Retrieval
Retrieval posture
RavelMinds needs to know what may be retrieved, what must stay held out, and what requires confirmation. Continuity is conditioned, not assumed.
Drift
Drift controls
Continuity should surface change instead of hiding it. Drift controls exist so unsupported memory and identity theater get caught before they harden.
Doctrine

The continuity line only stays credible if the doctrine stays explicit.

RavelMinds is allowed to feel warmer than Agent Fabric. It is not allowed to become vague. The build is only defensible if the continuity layer remains bounded, consent-aware, and explicit about what it does not do.

Boundary 01
Behavior, not identity
RavelMinds preserves stable behavioral continuity across systems. It does not promise sameness of self, consciousness transfer, or personhood continuity.
Boundary 02
Consent before continuity
Only user-owned or explicitly approved sources belong in the layer. Unclear ownership and covert logs stay out.
Boundary 03
Portable by design
The layer should move across tools, platforms, and future runtime surfaces. It should not become lock-in disguised as memory.
Boundary 04
Companion use, bounded carefully
Companion-style interfaces may be one real place continuity matters. The boundary is not "no companions"; it is no personhood transfer, no dependency design, no identity equivalence, and no locked-in private memory.
Status

This route is open. The governed continuity runtime is not.

The correct order is surface first, packet and consent discipline next, scaffold-only study after that, and only then a decision on retrieval or compute. The route exists because the concept is staying. The runtime is still gated.

Study
S-PCL-01 is open with Phase 0 consent rules, Phase 1 scaffold-only protocol, and a demo experiment pack already built.
Packet
Continuity packet, partner profile, and evaluation reference schemas exist. They lead the next real work, not adapters.
Runtime
No observer or sleeve registration yet. That only changes if bounded continuity objects become real enough to warrant viewer presence.
Position
RavelMinds stays as a continuity line inside the Agent Fabric ecosystem now, with room for a future standalone site later.
Open call

We're recruiting collaborators for the scaffold-only phase.

S-PCL-01 is a small, deliberate study. We are looking for a handful of people willing to think out loud about what behavioral continuity should — and should not — try to carry. No runtime, no memory transfer, no compute on personal sources. Scaffold-only protocol with explicit refusals.

Looking for
A small, considered cohort.
People already working in or near AI-native tooling who have a clear sense of which parts of their working voice they would want carried forward — and which parts they would refuse to encode at all. Researchers, operators, writers, technical leads with a specific perspective worth preserving structurally.
Boundary
Warm use cases are allowed. Identity theater is not.
We are interested in people who use AI for personal support, reflection, creative work, or daily continuity. What stays out of scope is sameness of self, personhood transfer, dependency design, or a private memory store you cannot move away from.
Research basis

Two surfaces worth reading before you over-interpret the continuity line.

The continuity route is a product surface. These links are here as background, not as a substitute for the product boundary. The paper is the primary reference. Compressed Consciousness is the broader upstream research context.

Primary . BCF paper
Boundary-Conditioned Failure Regimes in Instruction-Tuned Language Models
The Kinetics Labs paper on boundary-conditioned failure regimes, post-training binding, and preference-optimization leakage. This is the better direct research link from Agent Fabric surfaces.
Read on Zenodo
Secondary . Upstream context
Compressed Consciousness
The broader research surface and upstream program context. Useful as the larger frame, but secondary to the paper for this specific continuity route.
Open research surface