Deterministic Extraction
Extract any website into a normalized Universal Extraction Dataset. Headings, paragraphs, tables, images, links, metadata, JSON-LD — all captured without AI.
- DOM-based extraction
- Configurable recipes & profiles
- Multi-page crawling
Full-Text Search
Every extracted section, table, and metadata field is indexed with GIN-based full-text search. Find anything instantly across all your projects.
- GIN-indexed search
- Cross-project queries
- Section-level results
Multi-Format Export
Export your datasets to Markdown, JSON, or CSV. Your data is portable, machine-readable, and never locked in.
- Markdown with structure
- JSON with full fidelity
- CSV for spreadsheets
Universal Extraction Dataset
Every extraction produces a normalized record. Query sections, filter by type, join across pages. Your research becomes a database.
- Normalized schema
- Cross-page queries
- Statistics & metrics
Projects & Collections
Organize extractions into research projects. Create collections, tag findings, star important results, and build your knowledge base.
- Project-based organization
- Custom collections
- Tagging & starring
Optional AI Analysis
When you need deeper analysis, ask questions against your structured datasets. AI works on your normalized data, not in place of extraction.
- Query against datasets
- Context-aware answers
- Token usage tracking
Browser Extension
Extract any page with one click. The extension sends structured data to your workspace. No copy-paste, no manual formatting.
- One-click extraction
- Automatic structuring
- Workspace integration
Recipes & Automation
Define extraction recipes for different site types. Documentation, blogs, knowledge bases, e-commerce — each with optimized navigation rules.
- Pre-built recipes
- Custom configurations
- Automated navigation
Multi-Tenant Security
Organizations with role-based access. Row-level security on every table. Audit logs for all actions. SOC2-ready architecture.
- Organization workspaces
- Role-based access
- Full audit logging