Parinita  /  Products  /  Parinita Micro
Enterprise & Edge Layer

Parinita Micro

On-Premises AI Edge Endpoint

Sovereign AI at your location. No data leaves the facility.

Parinita Micro is an on-premises edge endpoint device that physically brings the Parinita AI platform's intelligence and security directly to a customer's location. It allows AI agents to operate locally in environments where data cannot leave the facility — hospital wards, factory floors, government field offices — with no cloud control plane required.

101
Sites Worldwide
$10
Per Device / Month
5
Max Per Seat
Air-Gap
Capable
Request a Demo → Explore Capabilities
01 / The Problem

AI Cannot Operate Where Data Cannot Leave

Hospital wards, factory floors, and government field offices run AI workloads that must process data on-site. Cloud-dependent AI inference violates data residency requirements. Traditional edge computing solutions do not include enterprise-grade security, PAM/IAM, or cryptographic attestation — they require separate software stacks installed and maintained by the customer.

Hardware-bound authentication for high-stakes operations (signing legal documents, releasing PHI, handling classified materials) requires physical proximity proof — a software token is not sufficient. No existing edge device provides a hardware anchor for trust elevation that integrates with a sovereign AI platform.

  • Cloud-Dependent AI in Air-Gapped Environments

    Cloud AI inference violates data residency for hospital wards, factory floors, and government field offices. Data cannot leave the facility but cloud AI requires network egress.

  • No Hardware-Bound Trust Elevation

    Software tokens cannot prove physical presence at a location. High-stakes operations (signing PHI release, accessing classified materials) require hardware proximity proof.

  • Separate Security Stack Per Device

    Traditional edge devices ship without enterprise security. Customers install and maintain Wazuh, PAM/IAM, and attestation separately — adding operational burden and inconsistency.

Capability
Current / legacy
What's needed
AI inference in air-gapped environments
Cloud AI — requires network egress, violates data residency
Qualcomm IQ-9075M NPU local inference — no data leaves the device or facility
Hardware-bound trust elevation
Software token — cannot prove physical presence
BLE or NFC tap on Micro device — hardware proof of physical proximity, elevates Presence to Bound or Sealed
Edge device security stack
Customer-installed Wazuh, PAM, attestation — per-device setup and maintenance
Factory-integrated Wazuh, Witness PAM/IAM, MCP Server, Chrysalis node — zero setup
PAM in air-gapped environments
Cloud-dependent PAM unavailable or degraded in air-gap
Witness PAM/IAM factory-integrated — operates with no cloud control plane
Edge device cost
Enterprise edge hardware at high per-device cost plus separate software licensing
$10/device/month — 5 devices per seat, Advantech-manufactured, all software factory-integrated
Agent tool execution at edge
No MCP server — agents cannot execute tools locally
Local MCP Server factory-integrated — agents (Forge, Guardian, etc.) execute tool calls on-device without cloud roundtrip
02 / Core Capabilities

Factory-integrated security. Local AI. Hardware-bound trust.

Parinita Micro ships factory-integrated with enterprise-grade software — Wazuh endpoint monitoring, Witness PAM/IAM hooks, a local MCP Server, and a Chrysalis attestation node. No software stack to install and maintain.

Two form factors: Micro (CP Edge Tiffin) — Qualcomm IQ-9075M SOM, Advantech-manufactured, $10/device/month, maximum 5 devices per Parinita seat. Micro Lite (P15) for lighter workloads.

  • Factory-Integrated Security

    Every Micro ships with Wazuh endpoint monitoring and Witness PAM/IAM hooks pre-installed. Zero security software to install and maintain.

  • Local MCP Server

    A local Model Context Protocol server allows AI agents (Forge, Guardian, and others) to execute tool calls directly on the device — no cloud roundtrip for edge-native actions.

  • Chrysalis Attestation Node

    Lightweight Chrysalis node on every device. Every session and local agent action cryptographically anchored and independently verifiable.

  • Local AI Inference

    Edge-native workloads run on the Qualcomm IQ-9075M NPU locally. Wake-word detection, local RAG, real-time vision — no cloud egress.

  • WireGuard POP Offload

    Heavier compute offloads to the nearest Parinita POP via secure WireGuard tunnel. Seamless local-to-POP handoff without reconfiguration.

  • Hardware-Bound Trust Elevation

    BLE or NFC tap proves physical proximity to the Micro device. Elevates Presence trust level to Bound or Sealed — required for signing legal documents, releasing PHI, or handling classified materials.

  • Air-Gapped Operation

    Witness PAM/IAM and Chrysalis attestation operate with no cloud control plane. Fully air-gapped environments supported on Micro (Tiffin) and Micro Lite (P15).

  • Pricing and Seat Equivalency

    $10/device/month. Maximum 5 Micro devices per Parinita seat. Five Micro devices consume the equivalent capacity of one platform seat.

  • Two Form Factors

    Micro (CP Edge Tiffin) — enterprise-grade, Qualcomm IQ-9075M SOM, Advantech-manufactured. Micro Lite (P15) — lighter workloads, lower profile.

03 / Architecture

How Micro Works

Micro serves two functional roles. First: local AI inference for edge-native workloads — wake-word detection, local RAG, real-time vision — while offloading heavier compute to the nearest Parinita POP via secure WireGuard tunnels. Second: hardware anchor for Presence trust elevation — BLE or NFC tap proves physical proximity and elevates the user's trust level to Bound or Sealed.

Every device runs a lightweight Chrysalis attestation node. Every device session and local agent action is cryptographically anchored and independently verifiable. Witness PAM/IAM hooks factory-integrated mean every privileged access event on the device is JIT-elevated and Chrysalis-anchored from day one.

P1
Reasoning Cortex
AMD Instinct MI350P
Primary AI inference · LLM serving
1,450+ nodes · 288GB HBM3e · high-bandwidth accelerator
P2
Training & Generation
NVIDIA RTX PRO 6000 Blackwell
Training · TTS · creative compute
950+ nodes · 96GB GDDR7
P3
Chain & CPU Compute
AMD EPYC Turin 9005
Chrysalis validators · CPU inference
700+ nodes · Zen 5c
P4
Knowledge & Retrieval
Intel Sierra Forest
Almanac vector search · RAG anchor
1,250+ nodes · 144 E-cores
P5
Long-Term Memory
NVMe Storage
Enclave · Stratum immutable object
850+ nodes · ransomware-resistant
P6
Media & Acceleration
RTX 4500 BSE · Alveo MA35D
Four tiers · GPU + FPGA + CPU
2,150+ nodes · 4K/8K hardware acceleration
P7
Edge Reflex
Qualcomm Cloud AI 100 Ultra
Ultra-low-latency edge inference
2,000+ nodes · sub-10ms response
P8
Coordination Layer
AmpereOne A128
Orchestra · Chorus routing · agents
2,400+ nodes · 128 ARM cores
P9
Nervous System
Cisco 8000 · Palo Alto · Arista
Routing · firewall · dual fabric
3,500+ devices · ConnectX-7 NICs
04 / Edge Compute Pipeline

Local inference. WireGuard offload. Chrysalis attestation.

Edge-native workloads run locally on the Qualcomm IQ-9075M NPU. Heavier workloads offload to the nearest Parinita POP via WireGuard tunnel. Every action — local or offloaded — anchored on Chrysalis via the device's attestation node.

  1. 01
    Device Boot
    Micro boots Instrument OS. Wazuh, Witness, MCP Server, and Chrysalis attestation node initialize.
  2. 02
    Network Connection
    Device connects to local network. WireGuard tunnel to nearest Parinita POP established automatically.
  3. 03
    Chrysalis Registration
    Device registers on Chrysalis attestation chain. Device identity anchored. Session begins.
  4. 04
    Local Workload Classification
    Edge-native workloads run locally on Qualcomm IQ-9075M NPU. Heavier workloads flagged for POP offload.
  5. 05
    POP Offload Decision
    Compute-intensive tasks offloaded to nearest POP via WireGuard tunnel. Local actions handled on-device.
  6. 06
    Presence Trust Elevation
    User BLE or NFC tap proves physical proximity. Presence trust level elevated to Bound or Sealed for high-stakes operations.
  7. 07
    Session Attestation
    Every device session and agent action anchored on Chrysalis via lightweight attestation node.
  8. 08
    Witness PAM Enforcement
    Privileged access on device handled by Witness — JIT elevation, session recording, Chrysalis-anchored revocation.
python 
# Micro ships ready to deploy:
# 1. Connect to local network
# 2. WireGuard tunnel to nearest POP established automatically
# 3. Chrysalis attestation node registers device identity
# 4. Wazuh + Witness active from boot
# 5. Local MCP Server available for agent tool calls
# 6. BLE/NFC available for Presence trust elevation
# No installation steps. No software to configure.
05 / Proof

Proven at scale. Not in a lab.

Parinita AI Edge is the production deployment of the Parinita platform and the largest heterogeneous AI infrastructure deployment in the United States.

Reference Deployment

Parinita AI Edge

The most complex heterogeneous AI infrastructure in the United States. 101 sites, 9 planes, 12,000+ nodes, 4 accelerator vendors, dual network fabrics, four-layer tenant isolation — all through a single sovereign control plane.

101
Points of Presence
4 tiers: T1 (32), T2 (29), T3 (19), T4 (21)
909+
K8s Clusters
101 sites x 9+ plane types
12K+
Compute Nodes
Supermicro, Dell, Ampere, Cachengo
4
Accelerator Vendors
Intel Habana, NVIDIA, AMD, Qualcomm

Network & Security Infrastructure

2,491+
Cisco Switches
+ 303 routers (EVPN-VXLAN)
1,734+
Arista Switches
Lossless GPU backend fabric
367+
Palo Alto Firewalls
PA-5580/PA-5560 series
152+
Petabytes Storage
NVMe over RDMA
  • Multi-vendor accelerators
    Four accelerator vendors — Intel Habana, NVIDIA, AMD, Qualcomm — orchestrated through one control plane with unified scheduling, monitoring, and lifecycle management.
  • Dual-fabric networking
    Cisco production fabric and Arista GPU backend fabric operating as a coordinated system, bridged by identity-aware routing.
  • Nationwide scale
    101 sites across 42 U.S. states, each operating autonomously with a local control agent and a sovereign cross-site routing plane.
  • Multi-tenant isolation
    Four-layer defense-in-depth: VXLAN VNIs, identity-routing, Palo Alto firewalls, and Cilium eBPF — validated across every plane and site.
  • Compliance readiness
    FIPS 140-2 at launch, with FedRAMP Moderate, CJIS, and IL4/IL5 certification paths active through Parinita compliance profiles.
  • Sub-millisecond routing
    Every request classified and dispatched in under 1ms, enabling real-time SLA enforcement without perceptible overhead.
06 / Use Cases

Built for environments where data cannot leave the facility.

Parinita Micro does not ask you to redesign your facility network. It is the sovereign AI endpoint that sits at your location — bringing the full Parinita platform to data that cannot move.

Hospital wards (HIPAA PHI that cannot leave the ward), factory floors (real-time vision and manufacturing AI with no external connectivity), government field offices (classified materials, air-gapped operations), and any regulated environment requiring on-site AI with hardware-bound authentication.

  • Healthcare Facilities

    Hospital wards where HIPAA PHI cannot leave the facility. Micro enables local AI inference (clinical decision support, patient monitoring) without data egress.

  • Manufacturing and Industrial

    Factory floors requiring real-time vision AI, anomaly detection, and manufacturing agent (Forge) tool execution — with no external connectivity.

  • Government Field Operations

    Government field offices and classified environments. Micro operates air-gapped with Witness PAM, Chrysalis attestation, and no cloud control plane.

  • High-Stakes Authentication

    Environments requiring hardware-bound authentication for document signing, PHI release, or classified material access. BLE/NFC proximity proof elevates Presence trust to Sealed.

  • Distributed Edge AI Fleets

    Enterprises deploying AI inference across dozens or hundreds of locations. Micro provides consistent, factory-integrated edge AI with centralized Chrysalis attestation across all devices.

07 / Getting Started

Deployment Options

Micro devices ship factory-integrated with all software. Deploy by connecting to your local network — the device establishes a secure WireGuard tunnel to the nearest Parinita POP automatically.

Model 1
Standard

$10/device/month — up to 5 devices per seat. Factory-integrated Wazuh, Witness, MCP Server, Chrysalis | Best for: Regulated facility deployment

Model 2
Air-Gapped

Micro Lite (P15) for air-gapped environments — no cloud control plane, full Witness PAM and Chrysalis attestation | Best for: Classified and air-gapped facilities

Full Scale
Fleet Deployment

Multi-site Micro fleet with centralized Chrysalis attestation, Parinita Central fleet management, and hardware-bound Presence trust elevation | Best for: Multi-site regulated enterprises

08 / Get Started

Request a Demo

Our engineering team has deployed Micro devices across regulated facilities. We bring that operational experience to every deployment conversation.

09 / Related

More from the fabric.

Products that compose with this one. Every Parinita product runs on the same 9-Plane Architecture across 101 edge POPs.