Learn · Research
RAG Chunk Planner
Estimate chunk count, vector storage, ingest time, and context-window pressure from corpus size and chunk settings.
Fill in your parameters below and click Calculate. All math runs locally in your browser — nothing is sent to a server or LLM.
Corpus
Chunking
Retrieval
After changing any field, click Recalculate to update results.
How we calculate
Corpus tokens are derived from pages × chars/page × tokens/char (or direct token/char input). Chunk count uses a sliding window: step = chunkSize × (1 − overlap%), then applies document-shape and strategy multipliers (heading-aware chunking reduces redundant splits; parent-child adds summary nodes).
Vector storage ≈ chunks × embeddingDim × 4 bytes (float32). Ingest time ≈ chunks ÷ embed QPS. Retrieval load = top-k × chunk size in tokens vs your generator context window.
Pricing assumptions as of 2026-08-30. Not a quote — verify vendor list prices.