Architecture overview
SENEX separates personal intelligence, governed exchange, and network coordination into distinct trust boundaries. This lets each layer be evaluated on its own instead of treating a future network as a prerequisite for useful local software.
The active foundation is AIA: local-first software for user-directed assistance, local context, and controlled interaction, implemented and tested for the scope in Current status. It is not a production release.
The system at a glance
1. User-controlled environment
AIA operates at the edge of the system, where the user's context and decisions originate. Its responsibility is to help the user while maintaining a clear boundary between local material and anything intentionally exchanged with another participant.
The owner's wallet and devices anchor this boundary: one wallet, a separate key for each device, and revocation per device.
The user-controlled environment is the primary trust boundary. Local possession does not, by itself, guarantee privacy: device security, operating-system behavior, backups, extensions, and user configuration also matter. SENEX therefore treats clear consent and observable data flow as product requirements rather than relying on the word "local."
2. Governed exchange
When cooperation is useful, the intended model is purpose-bound exchange. A participant should be able to understand what is requested, approve or reject it, limit the scope, and stop future access. Returned information should be bounded to the task rather than exposing an unrestricted source.
Exchange can be recursive: one owner's agent may ask another's, which may ask others, and each hop is governed by the owner of the data at that hop. The research goal is to send computation to the data rather than data to the computation. See Proof of Data.
The exchange layer is a capability boundary, not a promise that every collaboration technique is already implemented or formally private.
3. Coordination network
The Intelligent Chain is the planned coordination and accountability layer. It is intended to record accepted events, support consistent outcomes among participants, and provide evidence for testing contribution-aware processes.
The coordination network is a planned production-candidate test environment. It is not a live public blockchain. Test-only ASHA has no monetary value, and testnet participation creates no claim on V1.
4. Collective intelligence research
GENOME is the research direction concerning the collective intelligence that connected agents, governed data, computation, proofs, and outcomes may form. It builds on V1 foundations. It is not a deployed network mind, a current training system, or an implemented form of general intelligence.
Release boundaries
Components are built separately, but V1 is one integrated product; the SENEX documentation overview lists what it includes. V1-testnet and V1 are deliberately separate:
- V1-testnet: the production candidate, running the complete V1 product and protocol with test-only ASHA.
- V1: the planned production network, launched from a fresh genesis only after V1-testnet evidence and external review.
- GENOME: research considered after V1 foundations are stable; it is not part of the V1 launch claim.
Independent review is required before treating network integrity, privacy properties, availability, or contribution accounting as production-ready. Architecture diagrams and design intentions are not evidence of those properties.
Public documentation stays at the component, capability, and trust-boundary level. Internal protocols, thresholds, deployment topology, data schemas, model configuration, and security-sensitive operating details are intentionally excluded.