We aimed to demonstrate the feasibility of 90-day cardiac monitoring with an external Holter device and to find a target population able to benefit from such a technique. Cryptogenic stroke patients were continuously monitored for 90 days with a textile wearable Holter (TWH). Compliance and quality of the monitoring were assessed by the number of hours of ECG stored per month. Mean predictors of pAF, including age, gender, stroke severity, and atrial size (LAVI), were evaluated. One-year follow-up assessed pAF detection outside per protocol monitoring. Out of 224 patients included in 5 stroke centers, 163 patients (72.76%) fulfilled the criteria for the protocol. Median monitoring time was similar among the three months. Per protocol pAF detection reached 35.37% at 90 days. The age (OR 1.095; 95% CI 1.03–1.14) and the LAVI (OR 1.055; 95% CI 1.01–1.09) independently predicted pAF. The cut-off point of 70 years (AUC 0.68) (95% CI 0.60–0.76) predicted pAF with a sensitivity of 75.8% and specificity of 50.5%. The LAVI cut-off point of 28.5 (AUC 0.67) (95% CI 0.56–0.77) had a sensitivity of 63.6% and a specificity of 61.8% to detect pAF. The combination of both markers enhanced the validity of pAF detection sensitivity to 89.6%, with a specificity of 27.59%. These patients had increased risk of pAF during the 90-day monitoring HR 3.23 (χ² 7.15) and beyond 90 days (χ² 5.37). Intensive 90-days TWH monitoring detected a high percentage of pAF. However, a significant number of patients did not complete the monitoring. Patients older than 70 years and with enlarged left atria benefitted more from the protocol.