Foundation for AI Ethics & Safety Research

Demonstrating AI as a Tool
for Human Advancement

FAESR believes artificial intelligence should create abundance, not scarcity. We prove this through practical tools that advance human knowledge while remaining accessible to researchers, educators, and citizen scientists worldwide.

01

Practical Over Theoretical

We build working instruments, not position papers. Herald MSMO embodies our commitment to tangible scientific progress.

02

Radical Transparency

Open-source software, published methodologies, community governance. Every algorithm is auditable, every process documented.

03

Data Sovereignty

You own your observations. Choose Sovereign mode for complete local control, or Federated mode to contribute to open science—your data, your choice.

04

Abundance Mindset

Technology should multiply opportunities. As production scales, we pass savings to customers—not extract maximum margin.

Herald MSMO
Multi-Space Measurement Observatory

An automated space research station for citizen scientists. Herald monitors satellites, detects meteors, captures transient events, and times occultations— all from a single turnkey system. Software defines the mission. Real science, your roof.

Herald MSMO Multi-Space Measurement Observatory - concept render showing the three-camera system with Askar FRA400 telescope under a starry night sky
Concept render of Herald MSMO in the field. AI-generated image — final product appearance may vary.
3 Optical Sensors
BVRI Standardized Photometry
4 Science Modes
SDR Radio Ready

Three Cameras. Multiple Missions.

Wide-field detection triggers precision measurement. When Herald spots a satellite pass, meteor, or transient event, the AI pipeline coordinates all sensors to capture the science you need.

Photometry Ares-M Pro + BVRI Filters
Deep-Sky Mars-C II + 102mm APO
Trigger Ceres 462M Ultra-Wide
AI Detection → Classification → Alert

Software Defines the Mission

Same hardware, different modes. Tonight track satellites; tomorrow catch Perseids; next week time an asteroid occultation. All software is 100% open source— inspect every algorithm, modify anything, contribute back. Your observatory, your rules.

One Observatory. Four Missions.

Software-defined operation lets Herald switch between observation modes based on schedules, alerts, or your research priorities.

01

Satellite Tracking

Ingest TLEs, predict passes, capture light curves. BVRI photometry characterizes surface materials, tumble rates, and anomalies. Space situational awareness from your backyard.

SSA Light Curves Debris Tracking
02

Meteor Detection

Wide-field trigger camera catches events, photometry system measures standardized magnitudes. Shower association, fireball alerts, and radio correlation for 24/7 coverage.

IMO Fireball Networks Radio Meteor
03

Transient Events

Sprites, elves, and upper-atmosphere phenomena. High-speed capture on sensitive sensors catches millisecond events 50-90km up. Real science gap, citizen networks are sparse.

TLEs Sprites Space Weather
04

Occultation Timing

Star blinks out when asteroid passes in front. High-speed photometry nails precise timing. Coordinated networks measure asteroid shapes and discover moons.

IOTA Asteroid Science Predicted Events
05

BVRI Photometry

Standardized astronomical filters produce magnitudes directly comparable to professional observatories. Your data integrates with global databases across all modes.

AAVSO Research Grade All Modes
06

AI Pipeline

Detection, classification, alerting, and station adaptation. Learns your local conditions, correlates multi-sensor data, and routes observations to the right analysis path.

Auto Classification Smart Alerts

Software Architecture

Herald doesn't reinvent the wheel. We integrate battle-tested open-source engines under a unified scheduling layer that handles the hard problem: resource contention.

The Agent (Hypervisor)

Your camera can't run meteor detection and satellite tracking simultaneously. The Agent is a priority scheduler that manages this contention. It holds the master schedule, knows an ISS pass is coming at 04:15, pauses meteor detection at 04:14, hands the camera to the satellite tracker, captures the pass, then resumes meteor watch. No conflicts, no missed science.

Proven Engines

Each mission runs on established, community-vetted software. Meteor detection uses RMS (Global Meteor Network standard). Satellite prediction uses Skyfield/SGP4. Photometry uses SEP or AstroPhot. Radio meteor scatter uses Echoes. We orchestrate—we don't replace what already works.

Edge Layer (Fast)

A dedicated Raspberry Pi 5 with 16GB RAM and Hailo-8L AI accelerator (13 TOPS) runs lightweight trigger models at frame rate. "Something bright moved" or "RF spike detected" decisions happen in milliseconds. Dual NVMe storage buffers raw frames locally—what to keep gets decided after. This proven architecture powers 300+ scientific meteor stations worldwide.

Analysis Layer (Deep)

Heavy classification happens after capture. In Federated mode, events upload to cloud infrastructure for deep analysis and multi-station correlation. In Sovereign mode, raw FITS data streams to your local network—bring your own RTX 5090 or compute cluster for custom ML pipelines. The edge handles autonomous operation while power users push boundaries with experimental analysis.

Precision Engineering

Photometry System

Camera
ZWO Ares-M Pro (IMX533)
Sensor
9MP Mono, Cooled
Optics
Askar FRA 400 f/5.6
Filters
BVRI Photometric Set
Filter Wheel
ZWO EFW Mini (5-pos)
Output
Standardized Magnitudes

H-alpha Narrowband

Camera
Player One Ceres 462M (IMX462)
Sensor
2MP Mono, Uncooled
Optics
SVBONY SV106 60mm f/4
Focal Length
240mm
Filter
H-alpha 7nm Narrowband
Output
Hydrogen Emission Maps

Deep-Sky Imaging

Camera
ZWO Mars-C II (IMX662)
Sensor
2MP Color, Uncooled
Optics
102mm f/7 ED APO Refractor
Focal Length
700mm
Role
Dictates Pointing Direction
Output
Stacked DSO Color Images

Triple-Optics Architecture: All three cameras are co-aligned. Mars-C II targets a specific deep-sky object and dictates pointing. Ares captures BVRI photometry on stars in that field. Ceres captures continuous H-alpha emission data. Three complementary datasets from one pointing.

Electronics & Radio

Hub
Coolgear 7-Port Industrial
SDR Units
2× RTL-SDR Blog V4
RF Range
500kHz – 1.766GHz
Dew Control
3× Heater Bands + Controller
Connection
Single USB to PC
Frame
3D Printed ASA + Aluminum

Regional Note: European stations can use GRAVES radar (143.05 MHz) for radio meteor detection. The US lacks an equivalent always-on VHF source. We're evaluating alternatives including distant FM stations, NOAA weather radio, and aircraft beacons. Radio meteor capability in North America is experimental pending further testing.

Edge Computing

Processor
Raspberry Pi 5 (16GB)
AI Accelerator
Hailo-8L (13 TOPS)
Storage
Dual NVMe SSD Slots
Cooling
Active (Dual Fan + Heatsink)
Local Processing
Detection, Classification, Alerts
Power User Mode
Raw Data API for External GPU

Power Users: All raw FITS data accessible via local network API. Bring your own GPU rig for custom ML pipelines—the Pi handles autonomous operation while your hardware runs experimental analysis.

50 Stations. Three Ways to Join.

Building a geographically diverse observation network across USA & Canada. Buy one, sponsor one for a school, or apply for a sponsored station.

Sponsor a Station

Give One

Fund a station for a school

$ 7,999
  • You fund, a school receives
  • Your name on the station
  • Tax-deductible donation
  • Complete Herald MSMO system
  • Hand delivery & setup included
  • We match to qualifying institution
  • Updates from your sponsored station
  • Founding Network recognition
  • Put real science in classrooms
Sponsor a Station
Sponsored Station

Get One Free

For schools & research institutions

Free
  • No cost to your institution
  • Same hand delivery & setup
  • Curriculum integration support
  • Complete Herald MSMO system
  • All cameras, optics & filters
  • Tripod, case & accessories
  • 100% open-source software
  • Free Federated cloud forever
  • Commitment to regular use required
Apply for Sponsored Station
Territory: USA & Canada only for founding network
Site Diversity: All 50 stations curated for geographic and environmental variety
Future Pricing: Mass production pricing announced after successful network validation

Sovereign or Federated

Herald uses a split-brain architecture: time-critical decisions happen locally with zero latency, while deep analysis can leverage cloud compute. You control what stays on your machine.

Edge Processing (Always Local)

Trigger detection and frame buffering run entirely on your hardware. A meteor crosses your frame in 200ms—there's no time for round-trips. Local means instant.

Sovereign Mode

All processing local. Deep analysis runs on your GPU after capture. Data never leaves your machine. GPU requirements TBD after final optimization and testing.

Federated Mode

Edge detection local, deep analysis cloud-assisted. Multi-station correlation finds patterns no single observer could. Free forever. Contribute to open science.

Network Fusion

Last night, 47 stations captured data. Cloud processing correlates optical + radio across the network. Multi-modal validation eliminates false positives.

Your Choice

Switch anytime. Export everything. No lock-in. We believe your observations belong to you—always.

Modest Hardware Welcome

Federated mode offloads heavy analysis to the cloud. Edge detection runs lean. No high-end GPU required to contribute real science.

Join the Research Community

All Herald software is 100% open source. Transparent development, auditable algorithms, and a global network of citizen scientists advancing space observation together.

Frequently Asked Questions

What is Herald MSMO?

Herald MSMO (Multi-Space Measurement Observatory) is an automated space research station for citizen scientists. It tracks satellites, detects meteors, captures transient events like sprites, and times asteroid occultations—all from a single turnkey system with software-defined mission modes.

What is the difference between Sovereign and Federated mode?

Sovereign mode runs all processing locally on your hardware—data never leaves your machine. Federated mode keeps edge detection local for real-time response, but uploads captured events to cloud servers for deep analysis and multi-station correlation. Both modes use the same split-brain architecture where time-critical decisions (trigger detection) always happen locally.

What GPU do I need to run Herald MSMO?

None required. Herald includes a dedicated Raspberry Pi 5 (16GB) with Hailo-8L AI accelerator that handles all edge detection, classification, and autonomous operation. Federated mode uploads events to cloud for deep analysis. Power users wanting custom ML pipelines can tap raw FITS data via local network API and run their own GPU rig (RTX 5090, etc.) in parallel—the edge system keeps running independently.

Can I use Herald MSMO for radio meteor detection in North America?

European stations can use GRAVES radar (143.05 MHz) for radio meteor detection. The US lacks an equivalent always-on VHF source. FAESR is evaluating alternatives including distant FM stations, NOAA weather radio, and aircraft beacons. Radio meteor capability in North America is experimental pending further testing.

Is Herald MSMO software open source?

Yes, 100% of Herald software is open source. This includes the Agent (priority scheduler), all integrations with existing tools (RMS, Skyfield, SEP, AstroPhot, Echoes), and the edge/cloud processing pipeline. Every algorithm is auditable and you can modify anything.

When will Herald MSMO be available?

Founding Network launches Q2 2026 with 50 stations across USA and Canada. 25 Founders Edition units available for $7,999 with hand delivery and on-site setup. 25 sponsored stations donated to schools and research institutions through corporate and manufacturer partnerships. Mass production follows successful network validation.

Founding Network — Q2 2026

50 stations across USA & Canada. 25 Founders Edition for citizen scientists ($7,999 with hand delivery). 25 sponsored stations for schools and research institutions. All curated for site diversity—desert, coastal, high altitude, dark sky, and urban environments.

Citizen scientists, schools, research institutions, and potential sponsors welcome.