UPCOMING LAUNCH

Every flight leaves a trace. Read all your logs in minutes, not only when something goes wrong.

Multidrone automatically analyzes any ArduPilot .bin log — from routine maintenance flights to incident investigations. It reports how the aircraft behaved, detects anomalies and early degradation, and suggests actions. A technical assistance tool for your team, not a substitute for professional judgment.

Join the launch

Commercial product · No commitment · We'll notify you when available

Join waitlist →
THE MAIN CASE · DAILY USE

Operational audit after every flight

95% of flights end without incident, but every one of them leaves valuable information. The analyzer processes each log automatically and tells you what to look at, what's evolving, what's starting to drift. Early problem detection before they turn into accidents.

Nominal flight or not

Every flight gets an overall assessment: parameters within range, no relevant alerts, or specific observations to review.

OVERALL ASSESSMENT

Trends between flights

Vibrations growing flight by flight, batteries losing capacity, motors going out of balance. Spot it early and you fix it in the workshop, not in the field.

EARLY DEGRADATION

Time and operating hours

Accumulated hours, battery cycles, distance covered. Information ready for scheduled maintenance and for the reports your aviation authority requires.

OPERATIONAL LOG

Operational alerts

Failsafes triggered, manual mode switches, forced transitions, EKF lane switches. What the pilot should know about the flight even after a safe landing.

FLIGHT EVENTS
ANALYSIS GENERATED IN 23 SECONDS

Nominal flight · No relevant anomalies

QuadPlane (ArduPlane V4.6.3) completed a 22-minute inspection flight with a profile within operational parameters. VTOL → FW → VTOL transitions completed correctly. No failsafes, no critical alerts.

Flight summary

Duration22 min 14 s
Distance covered14.8 km
Maximum altitude120 m AGL
Average cruise speed19.2 m/s (69 km/h)
Battery consumption67% (of 22000 mAh)
Manual switch during flightNo
Failsafes triggeredNo

Observations

INFO · IMU 1 vibrations at 18.2 m/s² (nominal range <30). Trend stable relative to the previous 3 flights.
INFO · Difference between motors M1-M4 in hover: 4.2% (acceptable range <8%).
INFO · GPS HDop average 0.8 with 14-18 satellites. Fix quality nominal throughout the flight.

Recommendations

OK · No maintenance actions required at this time.
INFO · Next propeller inspection recommended at 8 accumulated flight hours (currently 47 h).

Note: Automated analysis of informational nature, generated from log data. Does not constitute an official appraisal. For critical operational decisions, validate with qualified technical personnel.
HOW IT WORKS

Three steps. Five minutes.

Upload the log you download from the autopilot. The system analyzes it completely: parameters, telemetry, events, vibrations, flight dynamics. You get a structured report ready to deliver.

01 · INPUT

Upload the .bin

Drag the log directly from your SD card or from Mission Planner. Up to 500 MB per file. Local processing — logs are not stored after analysis.

02 · ANALYSIS

Automatic diagnosis

The system cross-references thousands of events, parameters and messages to reconstruct the flight chronology. Detects anomalous signatures and provides technical evidence.

03 · REPORT

Technical report

Structured document with flight summary, observations, technical hypotheses and recommendations. Material for your technical team, verifiable and traceable.

THE EXTREME CASE · WHEN THINGS GO WRONG

Technical assistance to investigate incidents

When a flight ends badly, the log contains the key information of what happened. The system helps reconstruct the chronology and proposes technical hypotheses based on the data, along with reasoned discarding of alternative causes. Useful material for your team or your professional expert to advance much faster in the investigation.

Important: the system suggests hypotheses based on log data; it does not issue official appraisals or expert opinions with legal probative value. For insurance claims, judicial proceedings or official reports, findings must be reviewed, validated and signed by qualified personnel with the corresponding professional qualifications.

ILLUSTRATIVE EXAMPLE · BASED ON REAL CASES

Flight ended in accident · Technical hypotheses

This QuadPlane flight (ArduPlane V4.x, QuadPlane VTOL frame) ended in loss of control during RTL, with impact a few seconds later. The aircraft entered uncontrolled rotation with angular rates physically incompatible with an intact wing at cruise speed.

Flight chronology (time relative to arming)

ArmingT+00:00
Third RC Long Failsafe → RTLT+10:48
Angle assist triggered · roll ~−166° · pitch ~−29°T+10:56
Peak acceleration ~5.1 g · structural impact indicatorT+10:57
Abrupt end of logT+11:09

Quantitative physical analysis

The agent doesn't just read messages from the log. It cross-references sensor data (IMU, attitude, control, commands) to calculate derived magnitudes and contrast them with the physical limits of the aircraft model. Calculations that a human pilot would need hours to extract manually, and which are the basis for discarding or sustaining each hypothesis:

Angular acceleration ramp on roll axis> 200°/s² (rising)
Instantaneous roll rate> 200°/s
Transit time moderate bank → near inverted< 1 s
Peak vertical acceleration~5.1 g
Angular error DesRoll vs observed Roll> 140° deviation
IMU vibrations prior to eventWithin nominal range
Motor state (current and RPM)Nominal until the event

Main hypothesis suggested by the data

The combination of exponentially accelerating angular ramp with massive error between commanded roll and observed roll rules out autopilot command error: the aircraft does not respond to correction because it lacks the lifting surface to do so. Nominal vibrations and correct motor state also rule out propulsion mechanical failure.

Main hypothesis: acute separation of a lifting half-surface in flight (probable wing). Alternative hypotheses (aerodynamic stall, control surface servo failure, PID error, EKF degradation) are reasonably ruled out by the pattern observed in the quantitative data.

Secondary observations

ATTENTION · Chronic RC link degradation during flight (more than 20 link-loss events)
INFO · EKF3 lane switch detected after the event — consequence of impact, not cause

Important: automated analysis of informational nature based on patterns detected in the log data. Findings must be validated through physical inspection, comparison with other sources and review by qualified personnel. Does not constitute an official appraisal or report with legal probative value.
WHAT IT DETECTS

Beyond graphs and messages.

It's not a log viewer. The system reasons with knowledge of the firmware and the physical behavior of each type of aircraft. It gives you the technical material so you can make decisions; you sign and operate.

Temporal trends

Compare each flight with previous ones from the same drone. Detects gradual changes that are not visible in an isolated flight.

Precise chronology

Exact sequence of relevant events, from arming to landing, with precise timestamps.

Reasoned hypotheses

When there are anomalies, it proposes hypotheses based on the data and reasonably discards the alternatives.

Recommendations

Concrete actions: inspections to carry out, parameters to review, procedures to adjust. For your technical team.

ArduPilot knowledge

Reasons with firmware knowledge: parameters, Q_ASSIST, EKF, VTOL transitions, failsafes, flight modes.

Operational privacy

Your logs are yours. Not stored after analysis. Not used to train models. No third-party access.

WAITLIST

Let me know when ready

Join the Log Analyzer waitlist. We'll write to you when it's open to the public. No commitment, no newsletter, opt-out anytime.