The calculator amplified our power over arithmetic. Why has language been neglected by progress? Experience Cassandra, the source-anchored foundation for breakthrough research.
What you upload
Papers, notes, reports, manuals, transcripts, and mixed technical docs.
What Cassandra builds
A reusable Constellation Dataset that connects claims, methods, definitions, evidence, gaps, and citations.
What you keep
A source-anchored workspace you can chat with, inspect in tables, export, and reuse.
How It Works
Cassandra is built for the moment when search, folders, and one-off summaries stop being enough.
01
Explain your research goal in natural language, just as you would to your own research team.
02
Bring the messy research packet together instead of working one PDF or note at a time.
03
Review the size and build-time estimate, then approve your Constellation Dataset.
04
Ask source-anchored questions, explore the connected data visually, and export the result in portable formats.
Why LEGACHÉ Is Different
Documents do not become valuable by sitting next to each other. Cassandra enriches the corpus so the relationships between definitions, methods, claims, findings, and evidence become visible.
Drastically reduce friction from:
10x+
minimum target for richer data points after enrichment, giving a more complete corpus view.Scientific advisors
Cassandra is informed by advisors working across experimental optics, scientific machine learning, computational physics, and trustworthy AI.
Scientific Advisor
Experimental scientist at Bandelier Technologies working in quantum optics, photonics, sensing, and advanced instrumentation. He advises LEGACHÉ on fitting Cassandra into real scientific literature and cross-disciplinary research workflows.
Scientific Advisor
Computational scientist specializing in scientific machine learning, computational physics, and trustworthy AI. He advises LEGACHÉ on rigorous, interpretable systems for complex and consequential scientific work.
Use Cases
Compare claims, methods, findings, gaps, and citations across papers, then turn sources and notes into a reusable research workspace.
Analyze manuals, specifications, implementation documents, reports, methods, prior findings, and experimental context together.
Inspect evidence, definitions, stakeholders, and contradictions while anchoring claims, evidence trails, related work, and missing context.
Synthesize market, scientific, legal, and technical document packets into a source-anchored view.
Explore example corpora spanning biology, chemistry, data science, earth science, economics, psychology, and physics.
Accuracy at scale
Cassandra keeps smaller models focused on the right evidence as document volume grows, sustaining accuracy where model-only context starts to taper off.
| Documents | GPT Sol 5.6 | GPT 5.3 nano | GPT 5.3 nano + Cassandra |
|---|---|---|---|
| 10 | 96% | 56% | 94% |
| 100 | 92% | 50% | 94% |
| 200 | 84% | 43% | 93% |
| 300 | 74% | 36% | 92% |
| 400 | 64% | 30% | 91% |
Business Model
Choose a $25 one-time Constellation or an $80 monthly bundle with eight creations per paid billing period. Every Constellation remains hosted for life with unlimited chat sessions, lifetime ownership and downloads, and augmentations based on the option you choose.
Per constellation
One-time creation with lifetime hosted access, unlimited chat sessions, lifetime ownership and downloads, and two lifetime augmentations.
Monthly
Eight creations per paid billing period, with lifetime hosting, unlimited chat sessions, five lifetime augmentations each, and additional constellations for $25.