FlowMemory / Human Compute Network

The Human Compute Network.

FlowMemory turns phones, devices, agents, and verified work into a human-scale neural network for AI memory.

01 — Network

Phones become memory nodes.

HUMAN COMPUTE NODE LAYER

FLOW MEMORY AI MEMORY LAYER

human device signal receipt memory proof OWNER NODE ACTION VERIFIED STATE RECALL Phones · devices · agents · workers Signed receipts · proof roots · reusable memory HUMAN → COMPUTE SIX STAGES · ONE MEMORY NETWORK

Human Compute

The node layer that turns consented devices and agent actions into verifiable work signals.

Source: phones, devices, agents, workers, and live systems
Signal: signed receipts, verifier reports, policy trails
Output: proof-carrying work that can become AI memory

Flow Memory

The memory layer that compresses verified work into reusable context for AI systems.

Input: verified actions, receipts, outcomes, and memory roots
Capsule: compact state, trace pointers, proof roots
Recall: reconstruction paths for agents that need exact context

02 — Memory

A neural network for AI memory.

FlowMemory is neural in structure: many human-owned nodes, many signals, compact state, and feedback from verified outcomes.

The network grows as devices, agents, and work produce memories that can be checked, routed, and reused.

03 — Status

Built in layers.

Public primitives Hooks, warranted agents, receipt models, memory objects, and local conformance tools.
Device network Mobile workers, operator consoles, customer evidence, and verifier reports.
Launch gates Production custody, verifier networks, public deployment, and audited cryptography remain gated.

04 — Access