Your airline's workspace
Where your departments work with what your crews send from the Crewops app. Sign in with your airline's operator code: the products your airline holds open below. Fatigue (Crewops Insight), safety reports (Crewops Safety), training records (Crewops Academy) and controlled documents (Crewops Library).
ACCESS
Crewops Insight is licensed per operator. The same code that unlocks Crewops Crew for your pilots opens this page. No code yet? Run it on the sample roster, or ask for a trial on your own files.
Roster fatigue analysis
Drop a roster export, planned or flown: every duty against the FDP ceiling, predicted alertness through the duty (Three-Process Model, aircrew parameters), KSS and Samn-Perelli at top of descent, planned versus flown, and the post-flight ratings your crews share from the app. Files are read in this browser; nothing is uploaded.
PLANNED ROSTER
FLOWN ROSTER (optional)
OPERATION
STUDY ASSUMPTIONS · to set a run beside another model's study on the same terms
These change the run, not the roster. The assumptions used are printed with the results and in the CSV, so a difference from another study can be traced to its cause.
Summary
Planned versus flown
Duties
KSS 1–9 (Karolinska Sleepiness Scale) and Samn-Perelli 1–7 predicted at top of descent; EASA FTL 2.0 risk markers are KSS ≥ 7 and SP ≥ 5. Sleep is predicted from the roster, not measured. For augmented crews (three or more pilots listed) "EASA max" is the CS FTL.1.205(c) value for that crew and rest facility: 14/15/16 h with one additional pilot and 15/16/17 h with two (class 3/2/1), plus 1 h when one sector exceeds 9 h and the FDP has at most 2 sectors (c)(4); 3 sectors at most; 90 min rest each, 2 h for the landing pilots and "Rest" is the in-flight rest credited to this pilot as sleep, ending two hours before on-blocks; the operator's own ceiling prevails where an approved scheme sets it. "Night" follows ED Decision 2023/023/R: start 02:00–04:59 · end 02:00–05:59 · end after 06:00 (start before 02:00). PSWM is the GM5 CS FTL.1.205 prior-sleep/wake score (0 none · 1–4 low · 5–8 moderate · 9+ high).
Crew-reported, by destination
Samn-Perelli as the crew gave it after the flight, from the Crewops app, with the crew member's own switch on; the predicted figure is the model's SP at the worst landing of the same duty. The difference is reported minus predicted: positive means the crew felt sleepier than the model says.
By month
Model card
The full model card (model, equations, parameters, assumptions, verification status, limitations, references) is at hawkmason.com/crewops/insight/model-card; it accompanies every assessment report.
What this is. A roster reading with two layers: the duty-factor indicator Crewops shows crew (WOCL, early start, long FDP, sectors, short rest, consecutive duties, 7-day load, time zones; 0–100) and a biomathematical prediction from the Three-Process Model of alertness (Åkerstedt & Folkard; parameters re-estimated on 136 aircrews by Ingre et al. 2014, PLoS ONE 9(10):e108679), the same model family behind the Boeing Alertness Model. Sleep is generated from the duty pattern with the paper's thresholds; time-zone adaptation 30 % per day; KSS = 10.6 − 0.6 × alertness.
What it is not. Not a measurement of any person, not a validated commercial product, not a fitness-to-fly judgement and not an FRMS by itself. Per CASA (2010, 2014) and the IATA Human Factors Task Force (2025): model outputs support a risk assessment and compare rosters relatively; any threshold must be validated in your own operation; never a sole go/no-go. FTL tables are implemented from the published regulation and marked unverified in the Crewops app; your operator's scheme and system prevail.
Where the data goes. Nowhere. Parsing and modelling run in this page; close the tab and it is gone.
Safety reports
Reports your crews send from the Crewops app, with the flight, aircraft, stations and duty times filled in from their roster. Classify, act and reply here; the reply goes back to the reporter's phone. The 72-hour clock is Regulation (EU) 376/2014 Art. 4(8): a reportable occurrence reaches the authority within 72 hours; mark a report "not reportable" to stop it. The CSV follows the Annex I fields for your ECCAIRS or SHGM submission; it is an export, not a submission.
Training & qualifications
Read-and-sign notices land on every crew phone and come back confirmed, one tap each. Qualification records (OPC, line check, SEP, DG, CRM, cabin recurrent…) are uploaded as a CSV and shown to each crew member as their own; the validity periods follow ORO.FC.230 / ORO.CC.140 as your department sets them. Courses are assigned with a due date and a link; completion comes back from the phone for your department to verify. Records are kept for the periods of ORO.MLR.115.
EXPIRED AND EXPIRING · 30 DAYS
PUBLISH A NOTICE · READ AND SIGN
ASSIGN A COURSE
QUALIFICATION RECORDS · CSV
Columns: crew_code, role, name, kind, label, issued, expires, note, one row per item; a crew code's rows replace its record. Kinds: opc, lpc, line_check, sep, dg, crm, first_aid, route_competence, aerodrome_competence, recurrent_cabin, triennial_cabin, medical, licence, other.
NOTICES
COURSES
CREW RECORDS
Controlled documents
Manuals, procedures, bulletins and forms on every crew phone, current revision only. Publish a document once under its code (OM-A, CCM…); each new revision replaces the one on the phones, keeps the history here and removes the old offline copy. Mark a revision "read and confirm" and see who has confirmed it. The file stays on your own file host (an https link); the app downloads it once for offline reading.