Guides
Short, practical explainers organized by topic, each backed by the calculator, not generic filler.
Token basics
What tokens are and how counting works.
- What is a token? Pillar
The unit models read and bill on, explained plainly.
- How many tokens is 1000 words?
Tested word-to-token ratios by content type.
- How to count tokens in a PDF
Estimate a PDF token count before you send it.
Tools: Token counter · Tokens per word
Cost & pricing
How providers price input, output and cached tokens.
- LLM pricing explained Pillar
Per-token pricing, the input/output split and how to compare.
- Input, output and cached tokens
What each token type costs and when caching saves.
Tools: All model pricing · Token cost calculator · Compare models
Context windows
Fitting prompts and answers inside the limit.
- Context window vs max output Pillar
The two limits every request must satisfy.
Optimization
Cut tokens and cost without losing quality.
- How to reduce prompt tokens Pillar
Six practical techniques to cut input tokens.
- Token optimizer toolkit
The free tools that make up a practical optimizer.
- Prompt caching guide
How caching works and when the discount pays off.
Use cases
Cost models for common AI workloads.
- RAG cost guide Pillar
Why retrieved context dominates a RAG bill.
- AI chatbot cost guide
Turn users and messages into a monthly figure.
Tools: RAG cost calculator · AI chatbot cost calculator · AI agent cost calculator
Data formats
Tokenizing code, JSON, CSV and Markdown.
- Count tokens in a PDF
Estimate a PDF token count before sending.
Tools: Count tokens in JSON · Count tokens in code · Count tokens in CSV · Count tokens in Markdown