Free plan: 3 identifications and 1 follow-up per day, basic value range, up to 200 saved specimens when signed in. Plus plan: $2.50 per month billed yearly ($29.99 annually) or $9.99 per month billed monthly; 100 requests a day shared across identifications and follow-ups, detailed price analysis, m
Free: Unlimited messages with free models, access to Basic Models, real-time collaboration, side-by-side model comparison, 3 file uploads per day, 10 citation verifications, 10 GB Zotero Storage. Starter: $9 per month includes everything in Free plus Premium Models, 5x higher usage, 20 file uploads
Key features
Upload 1-3 photos of the same rock from different angles
AI-powered photo analysis to suggest rock matches
View visible clues, limitations and possible alternatives for each match
Follow-up questions about lookalikes and properties
Save photos, results and conversations to an account
Browse reference links to educational resources for suggested matches
Track daily identification allowance remaining
Price analysis and estimated value range for specimens
Search 200M+ academic papers
Verify citations against CrossRef and Semantic Scholar
Real-time collaboration with advisors and co-authors
Compare responses from multiple AI models
250+ integrations with research tools
Automated research workflows
Generate literature review tables
Create flashcards and study materials
PRISMA diagram generation
Cloud workspace to run code
What makes it different
No sign-up or app download required to start identifying rocks
Photo-based identification with transparent uncertainty acknowledgment
Works on any device with a browser and internet connection
Mobile-friendly with camera option where supported
All photo tools share the same subscription and daily allowance
Verifies every citation against academic databases instead of guessing
Addresses the citation problem where 1 in 4 citations misrepresent their source
Built on Open WebUI, the most widely used open-source AI interface
Real-time team collaboration in the same AI workspace
Cross-reference AI model outputs to catch hallucinations