CILoS is an indoor localization system based on CDMA mobile phone signal ﬁngerprinting. CDMA networks vary their transmission power to accommodate ﬂuctuations in network load. This affects signal intensity and therefore limits the practicality of traditional ﬁngerprintingapproachesbased on receiver signal strength (RSSI) measurements. Instead, CILoS uses ﬁngerprints of signal delay that are robust to cell resizing. We demonstrate that CILoS achieves a median accuracy of 5 meters, and compares favourably to RSSI ﬁngerprinting systems. We highlight the signiﬁcance of wide ﬁngerprints, constructed through scanning multiple channels, for achieving high localization accuracy. We also show that our system can accurately differentiate between ﬂoors of a multiﬂoor building.