About Tokenminning
Tokenminning is a community project for understanding, measuring, and governing LLM token usage in production systems. See tokenminning for the full definition.
The Manifesto summarizes the movement’s founding paper — the case for treating inference as a scarce resource. The Constitution turns that philosophy into engineering law: CI/CD gates, hard API ceilings, and immutable metering ledgers.
Why this exists
AI costs are dominated by token usage, yet most teams lack a shared vocabulary and reference for how that usage actually works. Tokenminning aims to be a practical, community-driven resource that anyone can read, share, and contribute to.
What we publish
- Manifesto: a paraphrased overview of the movement’s founding paper (source text not reproduced)
- Constitution: non-negotiable architectural primitives for production AI stacks
- Concepts: definitions and explainers on token economics, model selection, and cost attribution
- Practice guides: actionable techniques for minimizing token use in production
- Press: media coverage from The New York Times, TechCrunch, Fortune, and others — browse all press →
Start with Concepts for deeper dives into token economics, or Practice for actionable optimization techniques.
Built by Narev
This wiki is maintained by Narev , the platform for AI billing, cost tracking, and FinOps. Narev helps teams meter token usage, attribute costs to customers, and optimize model spend.
Contribute
Found something wrong or missing? Open an issue on GitHub .