Ani
A mesh-native personal assistance agent for individuals with ALS or severe mobility limitations — interpreting micro-inputs (eye-tracking, taps, gestures, minimal speech) and converting them into actionable, low-latency commands across the DOSFI and MeshInfer.AI ecosystem.
No IP Leakage · No Sensitive Information Exposure
MeshNativeExchange operates under strict no-IP-leakage and no sensitive information exposure rules. All contributor data is treated as confidential and ephemeral.
Contributor IP addresses are never collected, stored, displayed, or transmitted.
Device fingerprints, hardware IDs, and device metadata are never collected or stored.
Node metadata and routing information are never embedded in payout, verification, or dashboard output.
Contributor interactions are isolated to the current session. Data is discarded after verification unless legal retention is required.
Identity details beyond what the contributor explicitly provides during KYC are never inferred or revealed.
The verification workflow is designed to support GDPR data handling requirements. The SOC 2 Type II audit is a planned independent audit milestone and has not yet been completed. HIPAA-related controls apply when supporting regulated healthcare workloads.
Technical Capabilities
Overview
Ani is a personal assistance agent built on top of DOSFI and MeshInfer.AI, designed to provide real-time, adaptive assistance for individuals with ALS or severe mobility limitations. Ani interprets micro-inputs — eye-tracking, taps, gestures, and minimal speech — and converts them into actionable commands with ultra-low latency. The agent learns user patterns over time, adapts to weakening muscle input, predicts intent with minimal signals, and syncs across devices automatically. Ani operates even when offline, uses distributed inference for heavy tasks, and integrates with the MeshNativeExchange marketplace modules. Its interface is minimal, accessible, and mesh-native: large high-contrast buttons, eye-tracking compatible zones, one-tap emergency trigger, and adaptive layout based on user capability. Voice, gesture, and text are all optional input modalities.
Screenshots
Cross-device mesh sync — Ani operating across desktop, mobile, and browser nodes
Workflow Summary
Ani interprets micro-inputs (eye-tracking, taps, gestures, minimal speech) and routes them through intent detection, communication assistance, smart-home control, emergency signaling, daily routine automation, caregiver coordination, and device interoperability across mesh nodes.
State Machine
Ani maintains six states: Listening, Interpreting, Acting, Confirming, Escalating (emergencies), and Syncing (mesh nodes). Transitions are low-latency and optimized for ALS-friendly interaction.
Routing Policy
Local-first inference via MeshInfer.AI with nearby node acceleration and cloud fallback only when required. Prioritizes ultra-low latency for assistive actions and maintains privacy by keeping sensitive data local whenever possible.
Sandbox Preview
· Test safely, no infrastructure exposedSend a test input to preview how Ani responds. All execution is isolated — no real routing, storage, or DIU settlement occurs.
Version History
DOSFI Labs
View publisher profile →