Spectroscopy Analysis Platform
Upload your IR and UV-Vis spectra. Deterministic algorithms detect peaks with real Savitzky-Golay signal processing, then AI interprets the results — assigning functional groups and suggesting matching compounds with evidence-based confidence.
The demo loads a synthetic acetophenone spectrum (IR + UV-Vis) so you can explore the full workflow.
Five simple steps from raw data to interpreted results
Import CSV or TXT files with your IR or UV-Vis data. Files are stored privately — only you can access them.
Fine-tune the peak detection threshold. Real Savitzky-Golay signal processing identifies true peaks while filtering noise.
Deterministic range-matching assigns functional groups first, then AI writes the narrative and suggests matching compounds.
See peak assignments with evidence-based confidence, identified functional groups, and compound suggestions — labeled rule-based or AI-suggested.
Generate downloadable reports and export data in multiple formats for publication or further analysis.
Real Savitzky-Golay smoothing and derivative-based detection. Not a black box — every peak has a measurable prominence and width.
Functional groups are matched against a curated knowledge base first. Each assignment is labeled "rule-based" or "AI-suggested" with supporting evidence.
Uploaded spectra are stored privately and accessible only to you. Row-level security ensures no other user can see your data.
Your uploaded spectrum files and parsed data points are stored in a secure database with row-level security. Only you (and workspace admins) can access your data. No other user can view, edit, or delete your spectra.
When you run analysis, your detected peak data (positions, intensities, and widths — not the raw file) is sent to a large language model for interpretation. The LLM provider may temporarily process this data to generate the response. We do not control whether the provider retains this data for training — check the provider's policy for details.
Your spectra and analysis results remain in your account until you delete them. You can delete individual spectra at any time from the Analysis page, which also removes all associated analysis results.
If you are working with unpublished or sensitive research data, consider anonymizing sample names and removing identifying notes before upload. The peak positions and intensities themselves are not personally identifiable.
Try the demo with preloaded acetophenone data, or upload your own.