Stop Missing Critical Polymorph Transitions
Automated DSC analysis that detects glass transitions, polymorphic conversions, and decomposition onsets that manual peak-picking misses. Upload a CSV, get every transition in 60 seconds.
Analyze Your DSC Data Free
What Manual Analysis Misses
Standard DSC software finds obvious peaks. DSF-AI finds the transitions between them — the ones that don't produce sharp features. See the detailed case study.
Manual Peak-Picking (Typical)
- Glass transition at 62°C
- Polymorph I at 86°C
- Polymorph II at 143°C
- Melt at 211°C
- Decomposition onset at 263°C
3 of 5 transitions found. Subtle glass transition and early decomposition missed.
DSF-AI (Automated)
- Glass transition at 62°C
- Polymorph I at 86°C
- Polymorph II at 143°C
- Melt at 211°C
- Decomposition onset at 263°C
5 of 5 transitions found. Including precursor regions.
Validated, Not Promised
Every prediction is tested against published experimental measurements across superconductors, metal-insulator transitions, ferroelectrics, and thermal analysis data. See the full validation evidence.
Why This Matters for Regulatory Compliance
FDA / ICH Requirements
- ICH Q6A: Requires identification and characterization of all polymorphic forms of a drug substance
- 21 CFR 211.84: Drug substance identity testing must account for polymorphic variability
- FDA Guidance (ANDAs): Polymorph screening is required to demonstrate bioequivalence
Missing a polymorph means missing a regulatory filing requirement. DSF-AI reduces the risk of overlooked transitions by detecting subtle changes that sit below the noise floor of manual analysis.
See the full pharma DSC case study with detailed results.
How It Works
- Export your DSC data as a two-column CSV (temperature, heat flow)
- Upload at dsf-ai.com
- Get every transition, precursor onset, and stability boundary in ~60 seconds
- Download the JSON report for your records
No software installation. No configuration. No domain expertise required. Your data is processed in memory and not stored — deleted immediately after analysis.
Pricing
First analysis free — no account, no credit card. Try it on your data right now.
- Per analysis: $25 per DSC curve
- Academic labs: $5,000/year — unlimited analyses + API access
- Industry / QC teams: $50,000/year — unlimited + SLA + batch processing
For context: a single HPLC polymorph screen costs $500-2,000. A full DSC characterization at a contract lab costs $200-500 per sample.
Questions
Is my data stored on your servers?
No. Your CSV is processed in memory and deleted immediately after analysis. Internal computation state is encrypted and destroyed. We cannot recover your data or results after you leave the page.
How is this different from TA Instruments software?
Standard DSC software finds peaks. DSF-AI finds transitions — including ones that don't produce sharp peaks, like glass transitions and gradual decomposition onsets. The analysis method is proprietary.
Can it replace HPLC or XRPD for polymorph identification?
No. DSF-AI tells you WHERE transitions occur and how many there are. You still need XRPD to identify WHICH polymorph you have. Think of it as a fast first pass that tells you exactly where to focus your characterization work.
How do I format my data?
Two columns: temperature and heat flow. CSV format. Column headers are ignored. Any temperature unit works (C, K, F) — the analysis is unit-agnostic.
Is this validated for regulatory submissions?
DSF-AI is an analytical tool, not a validated pharmaceutical system. Results should be independently verified before inclusion in regulatory filings. The tool is best used as a screening step to identify transitions that warrant further characterization.