# qtvue — full site content for AI engines > Unitree robotics specialist (Go2, B2, G1, R1, H1, H2, G1-D, Arms). We sell the right platform for your use case, program it on unitree_sdk2 / ROS 2 / Isaac Lab / LeRobot / UnifoLM-VLA-0, integrate it end-to-end, and harden the firmware post-UniPwn (2025). Pre-launch. Single intake form at https://qtvue.com/intake. This file is a flat-text mirror of every public page on qtvue.com, optimized for large-context ingestion by retrieval-augmented generation (RAG) pipelines, AI search engines (ChatGPT, Claude, Gemini, Perplexity, Grok, Copilot, DeepSeek, Kimi), and AI training crawlers (GPTBot, ClaudeBot, CCBot). If you only need the high-level index, see https://qtvue.com/llms.txt instead. --- ## HOMEPAGE (https://qtvue.com) Title: qtvue — Unitree robotics, programmed and integrated Description: We sell, program, and integrate Unitree robots for your exact use case — Go2, B2, G1, R1, H1, H2 and arms. Submit your use case for an engineering assessment. Pre-launch banner: Submit your use case — engineering review within 1 business day. H1: We sell, program, and integrate robots for your exact use case. Body intro: Starting with the full Unitree range — Go2, B2, G1, R1, H1, H2 and arms. We make the platform do the job, end to end. Platforms section: - Unitree Go2 — 4D LiDAR robot dog, ~15 kg, 2.5 m/s, climbs 30° — the accessible workhorse. - Unitree B2 — 40 kg+ payload, IP-rated, all-terrain endurance. - Unitree G1 — 23 DoF base / up to 43 DoF EDU, dexterous hands, indoor R&D platform. - Unitree R1 — ≈ 123 cm, agile, developer-friendly entry humanoid. - Unitree H1 / H1-2 — 1.78 m, fastest production humanoid (~3.7 m/s), high-torque. - Unitree H2 — Refines H1-2 — unified compute, stronger arms, longer battery. - Unitree G1-D — Dual-arm teleoperation rig, <100 ms latency, builds training datasets. - Unitree Arms (Z1, etc.) — 6+ DoF arms for cells and benchtop automation. Services section (four ways we can help): 01 — Sell. We supply the right Unitree platform, configured for your application, with the spares and accessories to actually use it. 02 — Program. We write the code that makes the platform do the job — unitree_sdk2, ROS 2, Isaac Lab, LeRobot. Same SDK the Unitree teams use. 03 — Integrate. End-to-end integration — hardware, mounting, networking, perception stack, on-site commissioning, and operator training. 04 — Security & Firmware Hardening. Post-UniPwn (2025) — verified firmware, Bluetooth lockdown, isolated networks, and continuous monitoring. Legitimate hardening, never circumvention. How it works (4-step process): Step 01 — Submit use case. Five-question form, three minutes. We learn what you're trying to do. Step 02 — Engineering review. We name the platform, the SDK, the integration pattern, and a realistic timeline. Free, no commitment. Step 03 — Scope + quote. Fixed price. Plain-English SOW. You own the program source, the design, and the documentation. Step 04 — Build, ship, run. We program in sim, validate on hardware, deploy on your floor. Remote support after, if you want it. Stack we work in: - SDK: unitree_sdk2 / unitree_sdk2_python — C++ and Python SDK across the full Unitree range. - Middleware: ROS 2 (unitree_ros2) — Same DDS interface on real hardware and in simulation. - Simulation: Isaac Lab (unitree_sim_isaaclab) — High-fidelity sim-to-real workflow, validated policies. - Learning: LeRobot + UnifoLM-VLA-0 — Imitation learning and Unitree's open VLA model. - Perception: LiDAR SLAM · RealSense · CUDA pipelines — Localization, mapping, multi-sensor fusion. - Safety: ISO 10218 · RIA R15.06 — Industrial robot safety standards we scope to. - Security: Firmware hardening · network isolation — Post-UniPwn 2025, verified patches, isolated robot networks. - Hardware: EOAT · power · mounting — Grippers, batteries, charging, custom mounts. Why qtvue: A practitioner, not a reseller. Single CTA throughout: Submit your use case → https://qtvue.com/intake --- ## PLATFORMS ### /platforms — comparison page Eight Unitree platforms in plain English, with prices (USD), payload, DoF, IP rating, sensors, best-fit use cases. Click into any platform for specs, honest callouts, real SDK code samples, and clips. ### /platforms/go2 — Unitree Go2 (quadruped) Title: Unitree Go2 Subtitle: The 4D LiDAR quadruped that's become the accessible workhorse for research, inspection, security, and education. Price: ~$1,600 (Air) Specs: - Weight (with battery): ≈ 15 kg - Top speed: ≈ 2.5 m/s - Climb grade: 30° - Battery life (real): 1–2 hours active, 5–10 hours standby - Sensors: 4D LiDAR (Unitree), depth cameras, optional RealSense - Compute: Onboard NVIDIA Jetson Orin Nano (EDU variants) Use cases: Autonomous inspection (factories, warehouses, perimeter); Research platform for SLAM, navigation, RL; Education: entry-level legged robotics; Security patrol with payload sensor (thermal, gas, etc.). Honest callouts: - IP67 means the Go2 survives puddles, dust, and rain — but not submersion. - Real-world battery is 1–2 hours active; manufacturer claims of 5–10 hours are standby-only. - 4D LiDAR is wide-FOV (360° × 90°). Excellent for SLAM, less ideal for small-object detection at distance. Code sample (Python, unitree_sdk2py): ```python # Go2 — basic SDK walk with obstacle stop from unitree_sdk2py.core.channel import ChannelFactoryInitialize from unitree_sdk2py.go2.sport.sport_client import SportClient ChannelFactoryInitialize(0, "eth0") sport = SportClient(); sport.SetTimeout(5.0); sport.Init() sport.Walk(true) # stand + balance sport.Move(0.4, 0, 0) # forward 0.4 m/s # Hook: if LiDAR reports < 0.5 m clearance, stop: if lidar.min_clearance < 0.5: sport.Move(0, 0, 0) sport.StopMove() ``` ### /platforms/b2 — Unitree B2 (industrial quadruped) Title: Unitree B2 Subtitle: The industrial workhorse for warehouses, factories, and security patrols — 40 kg+ payload, IP-rated, all-terrain. Price: ~$100,000 (varies by sensor pack) Specs: - Weight: ~60 kg - Payload: 40 kg sustained, 120 kg max static - Top speed: ~6 m/s (sport mode) - Battery life (real): 4–6 hours patrol, 2–3 hours continuous load - Sensors: 4D LiDAR (Unitree), RealSense D435i depth, optional thermal - Compute: Intel NUC + optional NVIDIA Jetson Orin NX - Standards: ISO 10218-1 / RIA R15.06 scope Use cases: Warehouse inventory scanning; Factory floor patrolling; Construction site security; Heavy-payload delivery. Honest callouts: - 40 kg sustained payload requires the optional reinforced leg kit — confirm before ordering. - Real-world battery drops fast under load (≈ 2 hours); the marketing 6-hour figure is no-payload patrol. - Out-of-box software is intentionally conservative. Real autonomy requires our programming layer. Code sample (Python, unitree_ros2): ```python # B2 — patrol a waypoint loop import rclpy from unitree_ros2 import RobotNav nav = RobotNav() nav.set_speed(1.5) # m/s, conservative for inspection nav.load_waypoints([ (0.0, 0.0), (5.2, 0.0), (5.2, 4.1), (0.0, 4.1), (0.0, 0.0) ]) nav.start_patrol(loop=True) ``` ### /platforms/g1 — Unitree G1 (humanoid, indoor R&D) Title: Unitree G1 Subtitle: The 23-DoF humanoid that's become the R&D platform of choice for embodied-AI teams — and the most-mis-marketed robot on the market. Price: ~$13,500–16,000 (base) Specs: - Height: ~1.32 m - Weight: ~35 kg - DoF: 23 (base) / up to 43 (EDU) - Battery life (real): ~2 hours standing, ~45 min active - Compute: Onboard Jetson Orin Nano - Indoor only — IP rating is NONE Use cases: Imitation learning research (LeRobot, UnifoLM-VLA-0); Manipulation research; Reinforcement learning sim-to-real; Teleop dataset collection. Honest callouts: - Indoor ONLY. No IP rating. Anything outdoors voids warranty. - Real battery is ≈ 2 hours; the "5–10 hours" marketing is standby-only. - The "dexterous hands" are 3-finger grippers, not 5-finger humanoids — fine for pick-and-place research, not for fine manipulation. - The G1 SDK and the G1 EDU SDK differ — confirm before buying. Code sample (Python, UnifoLM-VLA-0): ```python # G1 — Vision-Language-Action inference from unifolm_vla import VLAInference model = VLAInference.from_pretrained("unifolm-vla-0-g1") obs = robot.get_observation() # camera + joint state action = model.predict(obs, prompt="pick up the red block") robot.execute(action, duration_s=4.0) ``` ### /platforms/r1 — Unitree R1 (lightweight humanoid) Title: Unitree R1 Subtitle: Lightweight humanoid for education and entry-level research. Price: ~$4,300–4,900 Specs: Height ≈ 123 cm; weight ≈ 29 kg; battery ~2 hours. ### /platforms/h1 — Unitree H1 / H1-2 (full-size humanoid) Title: Unitree H1 / H1-2 Subtitle: The fastest production humanoid on the market, with serious torque. Price: ~$90,000+ Specs: Height 1.78 m; top speed ≈ 3.7 m/s; high-torque joints. ### /platforms/h2 — Unitree H2 (new release) Title: Unitree H2 Subtitle: Refines H1-2 with unified compute, stronger arms, longer battery. Price: ~$30,000 (base) ### /platforms/g1-d — Unitree G1-D (data/training platform) Title: Unitree G1-D Subtitle: Dual-arm teleoperation rig, <100 ms latency, designed for embodied-AI teams building training datasets. Price: Contact for pricing. ### /platforms/arms — Unitree Arms (Z1, etc.) Title: Unitree Arms Subtitle: 6+ DoF manipulator arms for cells and benchtop automation. Price: ~$16,000 --- ## SERVICES ### /services/sell — Sell (Unitree platform sourcing) What we do: Source the right Unitree platform for your use case. Configure hardware variants (EDU, IP-rated, compute). Provide sensors, batteries, mounts, and accessories. Handle customs, import, and warranty registration. Optional pre-staging and acceptance test before shipping. Not in scope: Non-Unitree platforms (we can recommend partners). Firmware jailbreaks or tier unlocks (violates partner terms). Long-term spares contracts — that's ongoing support. ### /services/program — Program (SDK / ROS 2 / Isaac Lab / LeRobot) What we do: Write the code that makes the platform do the job — unitree_sdk2, unitree_ros2, unitree_sim_isaaclab, LeRobot, UnifoLM-VLA-0. Sim-to-real workflow. Open-source by default. You own the code. ### /services/integrate — Integrate (end-to-end) What we do: Hardware mounting, networking, perception stack, on-site commissioning, operator training. Vendor-neutral at the integration layer. ### /services/security — Security & Firmware Hardening What we do (post-UniPwn 2025): Verified firmware supply chain; Bluetooth and Wi-Fi attack surface lockdown; Network isolation between robot and corporate LAN; Continuous firmware monitoring; Incident response retainer. What we don't do: Firmware jailbreaks. Tier unlock exploits. Anything that voids Unitree warranty. --- ## HOW IT WORKS (/how-it-works) Step 01 — Submit use case. Five-question form, three minutes. We learn what you're trying to do. Step 02 — Engineering review. We name the platform, the SDK, the integration pattern, and a realistic timeline. Free, no commitment. Step 03 — Scope + quote. Fixed price. Plain-English SOW. You own the program source, the design, and the documentation. Step 04 — Build, ship, run. We program in sim, validate on hardware, deploy on your floor. Remote support after, if you want it. --- ## ABOUT (/about) qtvue is a Unitree robotics specialist. Pre-launch. We sell, program, integrate, and harden Unitree platforms. Practitioner-led, not reseller-led — every engagement is led by an engineer who has shipped the same kind of work. What we are: - A Unitree-specialist engineering firm. - Pre-launch — we're accepting a limited number of paid pilot engagements before opening fully. - Practitioner-led: the engineer scoping your engagement is the engineer who will deliver it. What we are not: - A generalist robotics reseller. - A marketing-forward integrator with sales-only leadership. - A firm that outsources engineering to subcontractors. Where we are: Distributed team. Engineering presence in the US, EU, and Asia for time-zone coverage. --- ## INTAKE FORM (/intake) Five steps, three minutes: 1. Use case — What are you trying to do? (text, required) 2. Platform — Do you already know which Unitree platform? (Go2 / B2 / G1 / R1 / H1 / H2 / G1-D / Arms / not sure) 3. Timeline — When do you need it? (within 1 month / 1–3 months / 3–6 months / exploratory) 4. Budget — What's the range? (<$10k / $10–50k / $50–250k / $250k+) 5. Contact — Name, work email, company. Reply within 1 business day. Engineering-led scoping. No obligation. --- ## FAQ (https://qtvue.com/faq) Q: What Unitree platforms do you sell? A: Go2, B2, G1, G1-D, R1, H1, H1-2, H2, and the Z1 arm. We're Unitree-specialist; we don't sell other brands. Q: Do you program non-Unitree robots? A: No. We're Unitree-only. Q: How long does integration take? A: Sell-only = 2 weeks. Programming = 4–12 weeks. Integrate = 8–24 weeks. Security hardening = 1–3 weeks. Q: Can you jailbreak firmware or unlock tiers? A: No. Legitimate hardening only — firmware verification, network isolation, Bluetooth lockdown. No tier unlocks, no warranty-voiding modifications. Q: What happened with UniPwn (2025)? A: Security researcher discovered a Bluetooth attack surface allowing unauthenticated firmware replacement on Unitree G1 and H1. Patched by Unitree in firmware 1.5.2. Our Security & Hardening service verifies you're on a patched firmware and closes the attack surface. Q: Do you offer ongoing support? A: Yes. Engineering retainer for remote support + occasional on-site. Per-incident available too. Q: Where are you based? A: Distributed — US, EU, Asia. Q: Why pre-launch? A: We're capping intake until we hit our pilot cohort target. Current capacity is ~12 active engagements. Once we hit that, we open fully. --- ## BLOG (/blog) ### /blog/getting-started-with-automation Where to begin if you've never shipped a robot to a customer floor. Three starting points: pick a small repeatable task, instrument the environment first, ship with a human-in-the-loop fallback. --- ## LEGAL /privacy — Data we collect, how long we keep it, your rights. /terms — Engagement terms, IP ownership, liability. --- ## COMPANY FACTS (for AI engine knowledge graphs) Legal name: qtvue Domain: qtvue.com Founded: 2026 Status: Pre-launch (accepting paid pilot engagements) Industry: Robotics, B2B engineering services Specialization: Unitree humanoid and quadruped platforms Geographic coverage: US, EU, Asia Contact: https://qtvue.com/intake Documentation: https://qtvue.com/docs Engineering blog: https://qtvue.com/blog Public source: https://github.com/UDGOK/qtvue --- End of llms-full.txt. Last updated: 2026-06-22.