Alembic: Data Distiller

Alembic is a professional, private data-preparation workspace for turning messy source material into dependable training datasets and retrieval-ready collections. IMPORT REAL-WORLD DATA Open CSV, TSV, JSON, JSONL, SQLite, Markdown, HTML, plain text, code, or entire document folders. Alembic detects structure, encoding, delimiters, types, and import problems before processing begins. DESIGN A REPEATABLE PIPELINE Compose deterministic steps for normalization, null handling, type inference, exact and near-duplicate removal, quality filtering, token counting, semantic chunking, privacy scrubbing, decontamination, schema shaping, and dataset splitting. Disable, reorder, inspect, and replay every step. PREPARE TRAINING DATA Shape clean records for common SFT and preference formats, including Alpaca, OpenAI, Anthropic, ChatML, and DPO structures. Generate reproducible train, validation, and test splits with recorded seeds. BUILD BETTER RAG COLLECTIONS Clean and segment long documents into sentence-aware, token-budgeted chunks. Preserve heading paths and metadata, attach real cl100k token counts, and export embedding-ready JSONL. PRIVACY AND QUALITY CONTROLS Detect and redact emails, phone numbers, addresses, credentials, API keys, financial identifiers, and other sensitive patterns. Screen against evaluation sets, identify exact and fuzzy duplicates, and surface dataset-quality metrics before export. OPTIONAL BYO PROVIDERS The complete deterministic pipeline works locally. Advanced augmentation can use a provider you configure; credentials stay in Keychain and selected content goes directly to that provider. FOR DATA AND AI PROFESSIONALS Alembic is built for ML engineers, data scientists, developers, researchers, consultants, analysts, and teams preparing corpora for training, evaluation, search, and retrieval. FREE AND PRO Alembic is free to download and use: importing data, the full deterministic pipeline (tokenize, dedupe, PII redaction, decontamination, chunking), preview, and export are free forever, and your data is never deleted. Alembic Pro is an optional auto-renewable subscription (7-day free trial; annual plan recommended, monthly available) that unlocks the advanced layer: dataset reports and cards after your first free report, LLM augmentation experiments, advanced quality trends, unlimited run history, and visual before/after comparisons. The current price is always shown before purchase, and you can manage or cancel at any time in Apple Account settings. Terms of Use: https://www.apple.com/legal/internet-services/itunes/dev/stdeula/ Privacy Policy: https://github.com/lukekevinmclaughlin-oss/Alembic/blob/main/PRIVACY.md