Background: Triage is routinely used to identify the patient’s level of emergency in crisis services. The SIDAS-FR was used for this purpose as part of the Quebec digital suicide prevention strategy. This study aimed 1) to evaluate the psychometric properties of the SIDAS-FR in an ecological context and 2) to assess its predictive validity as a triage risk screening tool for users of a suicide prevention chat service. Methods: During the triage of the suicide prevention chat service, the SIDAS-FR was systematically proposed to users, along with a few questions on age, gender, and risk of suicidal attempt. We collected triage data from N = 1278 users who completed the SIDAS-FR and participated in a crisis intervention (mean duration = 98 min) by a trained suicide risk prevention worker. The predictive validity of the SIDAS-FR was determined from the counselors’ estimate of suicidal risk at the end of the intervention. Results: A confirmatory factor analysis using the maximum likelihood estimation method replicated the unifactorial structure of the SIDAS-FR. All items showed satisfactory factor loadings, except for item 2 which was particularly low. The internal consistency of the instrument was good (α = .81; ω = .83). A ROC curve analysis showed a low sensitivity and low specificity of the SIDAS-FR to discriminate between individuals estimated to be at moderate or high risk of suicidal attempt. Discussion: As expected, the psychometric properties of the SIDAS-FR are weaker in an ecological setting. It is likely that the low factor loading on item 2 is caused by the use of a reverse-scored item. The use of reverse-keyed items is often criticized due to the difficulties of comprehension that it causes. This phenomenon could be even stronger in people in a state of crisis, likely to have impaired cognitive abilities. The benefits and limitations of using a cut-point for triage will be raised in the discussion.