Software
Tools built by pilots.
Powered by Maggie.
From raw drone footage to photorealistic 3D models. From waypoint planning to live AI monitoring. XFLIGHT software is built for the field — not the demo room.
XFLIGHT Flight Control
Autonomous flight. No GPS. No connection required.
Built on the battle-tested control loops of Betaflight, iNav, and ArduPilot, XFLIGHT Flight Control fuses proven open-source firmware logic with NVIDIA Jetson edge compute and live 3D Gaussian Splat maps. Drones localize themselves visually against photorealistic scene representations — maintaining full autonomy even when GPS is unavailable and Mission Control goes silent. Swarms self-organize into a mesh network, share sensor data cooperatively, and receive near-real-time map updates streamed from GPU Droplets trained by XFLIGHTScout.
Open-Source Firmware Core
PID controllers, flight modes, and mixer logic derived from Betaflight, iNav, and ArduPilot — validated on thousands of airframes, extended for swarm use cases
GPS-Free Visual Navigation
3D Gaussian Splat maps replace GPS as primary localization: 6-DoF pose estimation from live camera feed matched against pre-built splat scenes in real time
NVIDIA Jetson Edge Compute
Jetson Orin NX runs neural odometry, monocular depth estimation, and obstacle detection on-board — 30+ fps, zero cloud latency
Swarm Mesh Network
Every drone is a mesh node. On Mission Control link loss, the swarm self-organizes, re-routes telemetry and map data through peers, and continues the mission cooperatively
Live Gaussian Map Updates
XFLIGHTScout streams freshly trained splat tiles from DigitalOcean GPU Droplets to the swarm in near real-time — navigation maps update mid-flight as new terrain is scanned
Connection-Loss Resilience
Autonomous fallback activates instantly on link loss — last known mission plan continues, swarm consensus maintains formation integrity, data is synced on reconnect
XFLIGHT Scout
Drone footage becomes a 3D world.
Feed it a GoPro flight. Get a fully navigable, photorealistic 3D model. XFLIGHTScout turns raw drone video into Gaussian Splat scenes you can explore in real time — running entirely on your Mac.
GoPro Video Ingestion
GPMF metadata extraction, automatic flight segment detection
COLMAP SfM Pipeline
Structure-from-Motion with OPENCV_FISHEYE — optimized for drone optics
3D Gaussian Splatting
OpenSplat training with Metal GPU acceleration — 60+ fps playback
MapKit Integration
Geospatial project management with GPS-tagged 3D scenes
Apple Silicon Optimized
Full MPS pipeline — no external GPU required
Cloud Training (coming)
DigitalOcean GPU offload for high-resolution reconstructions
XFLIGHT DroneMapPro
Plan precise. Fly smart. Document everything.
Professional mission planning and post-flight analysis for inspection, survey, and autonomous operations. Built by pilots who fly in places most software has never heard of.
3D Mission Planning
Terrain-aware waypoint routing with obstacle avoidance integration
Inspection Mode
Automated flight paths for bridges, turbines, and infrastructure — pre-calculated, repeateable
Inspection Reports
Auto-generated PDF/HTML reports with geo-tagged images and anomaly markers
Airspace Overlays
Real-time regulatory zones, NOTAMs, and restricted airspace visualization
Live Telemetry
In-flight monitoring with Maggie AI integration for mission status
Scout Integration
Direct handoff to XFLIGHTScout for post-flight 3D reconstruction
Maggie Brain Visualizer
See exactly how Maggie thinks.
A real-time window into Maggie's knowledge graph. Watch memory vectors form, semantic clusters emerge, and connections grow — transparent AI, not a black box.
Vector Database View
Live inspection of Maggie's SQLite knowledge store with semantic clustering
Knowledge Graph
Visual connections between concepts, projects, and decisions
Memory Timeline
Chronological view of when Maggie learned what — and from which session
Semantic Search
Query the vector store visually — find related memories by concept proximity
Growth Analytics
Track how Maggie's knowledge base expands over time
100% Local
No data leaves the machine — all visualization runs on XFLIGHT hardware
Maggie Operations Panel
Every agent. Every mission. Real-time.
The nerve center for monitoring Maggie and all active OpenClaw agents. Live session states, event timelines, artifact tracking — everything XFLIGHT needs to know what's running and what just finished.
Live Session Dashboard
Active, waiting, warning, and done states — prioritized and always up to date
Event Timeline
Full event stream per session with timestamps and state transitions
Artifact Tracking
File output validation — knows which artifacts exist and which are still pending
Agent Relationships
Parent-child session graphs showing how Maggie delegates and coordinates
Warning Detection
Automatic surfacing of stalled sessions, failed agents, and resource alerts
DevForge Integration
Full visibility into DevForge build pipelines running under Maggie's coordination
One ecosystem. One intelligence.
XFLIGHTScout scans and trains Gaussian Maps — streamed live to Flight Control drones. DroneMapPro plans the missions. The swarm executes autonomously, monitored by the Operations Panel. The Brain Visualizer shows how Maggie learns from all of it.
Interested in early access or partnership?
Some products are available now. Others are in active development. We're open to research partners and enterprise pilots.
Get in touch →