Systems and methods provide the generation of signed recordings of audio and video calls. The signed recordings include information that may be used to verify the integrity of a purported copy of the recording. The recording may be generated by capturing the audio and/or video from a monitored communication and streaming the captured data to a recording. As the captured data is streamed to the media file used for the recording, a hash value is calculated for each segment of the received data. Information describing the context of the communication may be used to generate another hash value. A digital signature for the recorded call may be generated based on the two combined hash values. Operating in this manner, a signed recording of the communication may be generated while a recording of the communication to a media file is being made, rather than by post-call processing of a stored file.