The documentation clearly distinguishes between actual spine angle measurement and simple activity tracking. The local inference model appears reliable for basic ergonomic monitoring.
Questions about FIXPOSTURE / PRO
What is FIXPOSTURE / PRO?
FIXPOSTURE / PRO is an edge AI biomechanics engine that performs real-time posture and spinal health monitoring using on-device computer vision. It analyzes video from a standard webcam at 60 frames per second to measure cervical, thoracic, and lumbar biomechanics locally without uploading any video data. The system tracks metrics such as C1–C7 equivalent load, L1–L5 compression, muscle strain index, and composite posture health scores. It provides a live simulator, benchmark tools, and an SDK, and is currently offered as a private beta with invite-only whitelist access.
How does the biomechanics engine process data?
The engine runs entirely on-device with a 33-keypoint pose graph lifted into a sagittal biomechanical model to solve the moment arm at each vertebral level. This approach provides load in kilograms rather than a vague flag, and it operates at 60 Hz continuous sample rates. The system is HIPAA and GDPR compliant by architecture, not by policy, ensuring privacy through local inference only.
What is the Micro-Nudge UI feature?
The system delivers correction as a 2 px luminous gradient at the edge of your display, calibrated below the threshold of conscious interruption. This micro-nudge form avoids modals, sound, or dopamine loops to keep the flow state intact. The adaptation uses a per-user habituation model to prevent over-correction.
What are the pricing plans and access seats?
Solo plans cost $18 per month and include 1 seat with 2 devices, providing full engine access with no telemetry. Studio Cohort 03 costs $64 per month for 10 seats with unlimited devices, adding shared calibration profiles and CLI access. Enterprise plans offer custom unlimited seats with on-prem model registry and signed builds for physical-therapy or occupational-health orgs.
How do I request access and what are the terms?
To get started, you must request whitelist access because seats are released in cohorts rather than sold openly. Each seat is paired with a reviewed anthropometric profile, so the system releases seats in batches to maintain calibration integrity. You can cancel mid-cycle as seats are month-to-month and return to the cohort queue immediately.
What hardware is required to run the software?
The system requires any 720p webcam and a machine from the last six years, with no wearable, strap, chair sensor, or external hub needed. Inference runs entirely in the local runtime using WebGPU / CoreML / DirectML with a frame latency of 6.2 ms @ 1080p. The engine emits joint angles and load scores but never uploads frames, crops, or face embeddings.
How does the live simulator work?
The live simulator allows users to drag biomechanical parameters to watch cervical and lumbar load respond in real time. This exact model runs on-device at 60 Hz and is driven by sliders instead of a camera for demonstration purposes. It visualizes sagittal wireframes and local inference results without needing a live feed.
How does the benchmark matrix compare to other products?
The benchmark matrix compares FIXPOSTURE / PRO against posture braces, standing desk timers, and wrist wearables. While braces restrain and timers nag, this system uses a calibrated camera to measure sagittal alignment and dwell time. The benchmark highlights the difference between accurate but cloud-bound tools and private but useless ones.
What specific metrics does the system track?
The engine solves the moment arm at each vertebral level to provide specific load measurements in kilograms. It tracks C1–C7 equivalent load and L1–L5 compression alongside muscle strain index and composite posture health scores. These metrics are calculated continuously at 60 Hz to monitor spinal health in real time.
What is the current status and location of the company?
The system is currently in private beta with invite-only whitelist access, meaning it is not yet publicly available for purchase. The company was established in 2026 and is located in Palo Alto, CA. The engine version currently running is v2.6.0 with local inference only.
What do reviewers praise and complain about?
Some reviewers on Trustburn praise the local-only inference that keeps frames entirely on the device. Others complain about the invite-only whitelist system which limits immediate access to the engine. The overall rating is 3.4 out of 5 from 5 reviews, reflecting mixed experiences with the beta access model.
Who is this system best suited for?
The system is designed for people who sit for a living and need clinical-grade spinal sensor capabilities without external hardware. It suits users who want accurate biomechanics data without cloud dependency or wearable devices. It does not suit those who need immediate public access without waiting for a cohort release.
What development tools and SDKs are available?
The engine provides a TypeScript SDK and CLI access for developers who want to script against the runtime. Enterprise users get an on-prem model registry and signed builds for deployment across physical-therapy or occupational-health orgs. The SDK allows deep integration with the local-only engine for custom joint-load models per clinical protocol.
What are the performance specifications?
The system operates with zero hardware requirements, using only a standard webcam and existing computer resources. It processes frames at 60 frames per second with a 6.2 ms frame latency at 1080p resolution. The inference target supports WebGPU / CoreML / DirectML for cross-platform compatibility.
How does the system ensure data privacy and compliance?
The system is HIPAA and GDPR compliant by architecture, ensuring data privacy through local processing. It runs air-gapped with no network egress, meaning the tensor never touches a socket and the buffer is discarded after inference. This design ensures that no video data leaves the device during operation.
AI-generated from reviews on this page. Verify anything important with the company directly.
The edge inference model looks promising for remote workers who value data sovereignty. With zero bytes leaving the device and no cloud dependency, the HIPAA and GDPR compliance by architecture rather than just policy is a strong differentiator. The 6.2 ms latency at 1080p suggests a smooth experience. I appreciate that the micro-nudge stays below the conscious threshold, preserving flow state without intrusive modals or sound. It feels like a mature product for its private beta stage.
The SDK reference code shows that importing the engine is straightforward, allowing for ergonomic-aware apps with minimal boilerplate. The ability to ship the engine inside Swift, Electron, or Web products is a huge plus for developers. I like that the enterprise licenses include the raw keypoint stream and biomechanics solver, which adds significant value. The hardware-free approach means users do not need to wear a brace or wristband, reducing the risk of muscle deconditioning. The micro-nudge output seems unobtrusive, calibrated to stay below the threshold of conscious interruption. This could help maintain focus while still providing real-time feedback on cervical and lumbar load. Overall, the combination of privacy, precision, and ease of integration makes this a strong contender in the ergonomic software space.
The per-seat whitelist gating is clear, though I wonder if the calibration fee is included in the enterprise license cost for the biomechanics solver.
Reading through the biomechanics specifications, the approach to measuring vertebral load in kilograms rather than just vague posture flags is genuinely impressive. The fact that the engine runs locally on WebGPU, CoreML, or DirectML means there is no network egress, which solves the privacy concerns that usually plague cloud-based posture tools. The 33-keypoint pose graph lifted into a sagittal model provides a level of detail that wearables simply cannot match. I am particularly intrigued by the micro-nudge UI, which uses a 2 px luminous gradient at the display edge. This seems designed to correct alignment without breaking concentration, as it adapts to per-user habituation. The setup time of just thirty-eight seconds compared to five minutes for standing desks is also a significant advantage for daily use. However, since it is currently in a private beta with whitelist access, the full scope of features available to standard users remains to be seen. The SDK integration looks straightforward for developers, suggesting that third-party ergonomic apps could easily adopt this technology. For now, the focus on local inference and hardware-free operation makes it a compelling option for those who sit for long hours and want precise, private data about their spinal health.
News
No news yet. This is where the business shares its own updates and press.
Photos
No photos yet. Photos shared by reviewers will show up here.
Widgets
Get the widget for your site →About FIXPOSTURE / PRO
FIXPOSTURE / PRO is an edge AI biomechanics engine that performs real-time posture and spinal health monitoring using on-device computer vision. It analyzes video from a standard webcam to measure cervical, thoracic, and lumbar biomechanics locally at 60 frames per second, without uploading any video data. The system tracks metrics such as C1–C7 equivalent load, L1–L5 compression, muscle strain index, and composite posture health scores. It provides a live simulator, benchmark tools, and an SDK, and is currently offered as a private beta with invite-only whitelist access.
- Website
- fixposture.pro