Explore proprietary AI text research, the recovery and restoration of Sikh images, and more than 38,000 extracted kirtan, simran and katha audio links.
THE RESEARCH & RECOVERY PROGRAMMEDECODE, TRANSLATE, PRESERVE.
Recover the record. Ask better questions.
SikhArchive brings together proprietary AI research, English translation, visual recovery and audio discovery. The purpose is to make dispersed material readable, findable and available for further study.
01 / DECODING TEXTS & TESTING READINGS
Proprietary AI. Evidence-led enquiry.
The archive’s in-house AI work extends beyond translation into the investigation of difficult texts and possible readings. Decoding is a research question: what does a source say, what supports a proposed reading, and what remains uncertain?
For a proposed reading to become a research finding, it needs the source image, a transcription, an explanation of the method, comparison with alternatives and review. No decoding benchmark or decipherment result is asserted here.
SikhArchive’s gallery project brings long-lost and overlooked Sikh images back into circulation, reconnecting readers with photographs, paintings, portraits and visual records scattered across collections.
The AI restoration project sits alongside that work, showing damaged photographs and paintings before and after restoration. Finding a source and reconstructing missing detail are different parts of the process; consult the original alongside the restored image.
Kirtan, simran and katha: a substantial body of recorded Sikh sound made discoverable through extracted audio links. The collection supports listening, playlists and the 108 FM radio experience.
These are links to audio recordings, not a count of unique compositions, artists or newly recorded performances.
Research, image-recovery and audio-extraction descriptions supplied by SikhArchive. The sacred-music page independently reports 38,000+ tracks. Methods and results should be assessed at the individual source or project level.
