Abstract: A computer system for providing on-demand sign-language translation services receives a first audio recording of ambient noise. The computer system then accesses a database of audio fingerprints. Using the audio fingerprints, the computer system identifies within the first audio recording a particular audio fingerprint from the database of audio fingerprints. The computer system then identifies within the particular audio fingerprint a particular timestamp associated with the first audio recording. The computer system accesses a database of translated audio. The computer system accesses a portion of the translated audio that corresponds with the particular time stamp. The computer system then displays on a user interface the portion of the translated audio.