Pocket2 - Seamless Real-time Spatial Scanner

MindPalace

Enterprise power. Team-level decision.

MindPalace Pocket2 is a handheld, real-time spatial scanning productivity tool. It integrates high-precision depth sensing, real-time point cloud reconstruction, wireless data transmission, and cloud collaboration into a single, streamlined device.

Powered by Manifold Tech proprietary MindSLAM® algorithm, Pocket2 achieves hardware-level synchronization across multiple sensors and pixel-level data fusion, removing the traditional complexity of 3D reconstruction and making professional-grade scanning accessible even to non-experts. With built-in multi-scenario templates and one-click cloud upload, Pocket2 quickly generates standardized deliverables for a wide range of applications, including interior measurement & modeling, accident reconstruction, forestry surveys, digital asset capture, stockpile & volume calculations, and robot deployment.

Whether for individual users, collaborative teams, or industry-scale projects, Pocket2 simplifies complex workflows into a single, standardized process, enabling faster execution and higher efficiency

CORE FUNCTIONS

Multi-Modal Sensor Fusion

No training required – simply power on and scan. Pocket2 combines a global shutter camera, LiDAR, and IMU with the MindSLAM® algorithm to deliver stable odometry and real-time, high-precision colored point clouds, even in complex or dynamic environments.

Real-Time Point Cloud & Measurement

Pocket2 supports a true "scan and see" workflow, allowing instant detail refinement. This reduces sparse areas and occlusion blind spots while ensuring data completeness, accuracy, and usability for measurement and modeling.

One-Tap Cloud Sync & Multi-Device Collaboration

It connects seamlessly with your app and the cloud, automatically generating high-quality, lightweight 3D files. Data can be viewed, edited, and exported across devices, enabling real-time collaboration and accelerated project delivery.

Robotic Collaboration Ecosystem

More than a scanner! It is the entry point to the Manifold Tech ecosystem. Through deep integration with the Odin module, it feeds precise 3D environmental data into robotic systems, enabling advanced perception and autonomous decision-making.

TYPICAL APPLICATION SCENARIOS

Interior Measurement & Renovation Planning

With one-click cloud upload after capture. Pocket2 enables fast generation of CAD drawings and dimensioned files, providing solid data support and clear visualization for design, construction, and reporting.

Forestry Survey & Ecological Sampling

For forestry applications, Pocket2 labels tree species and locations while generating high-density forest point clouds. It supports exporting structured data such as species type and spatial distribution, making it easy to produce sampling files and statistical reports.

Digital Asset Capture & Cultural Heritage Recording

For cultural heritage and historical architecture, Pocket2 captures facades, textures, and structural details with high fidelity. The data can be imported into Blender, Unreal, or BIM platforms for modeling, restoration, and digital reconstruction.

Robotic Deployment & Navigation Mapping

Use Pocket2 to quickly scan robotic work environments and generate high-precision navigation maps with one click. When combined with the Odin module, it enables rapid robot deployment, accurate relocalization, and dynamic obstacle avoidance.

Technical Specifications

Dimensions - 4.5 × 4.4 × 4.7 in (115 × 113 × 120 mm)

Device Weight - 27.84 oz（790 g）

Battery Handle Weight - 17.12 oz（485 g）

Operating Temperature - 14°F to 104°F (-10°C to 40°C)

Data Interface - USB 3.0

Storage - Built-in SSD, 256 GB

Battery Capacity - 4000 mAh

Battery Life (Single Battery) - Up to 120 minutes

Point Cloud Accuracy - Better than 0.4 in (≤ 1 cm)

Camera System - 3 × 2 MP Global Shutter RGB Cameras

Laser Wavelength - 905 nm

Laser Channels - 40 channels

Point Rate - 200,000 points/sec

Measurement Range - 131 ft @ 10% reflectivity / 230 ft @ 80% reflectivity (40 m / 70 m)

Global Positioning Accuracy - Better than 2 in (≤ 5 cm), typical ~1.2 in (≈ 3 cm)

MindCloud Studio

• Output Formats: .las, .laz, .fbx, .e57, .pcd, .ply, .obj

• Supports USD export (NVIDIA Omniverse) for robotic simulation and training

• Generate grayscale and true-color mesh models

• Support external control point import

MindCloud Go（APP）

• Annotation mode with label import and quick review workflow

• One-tap switching between intensity and true-color rendering modes

• One-click cloud upload for LAS and colored point cloud data