About TokenTally
Built for teams who hate guessing their AI bill.
TokenTally is an independent project focused on one thing: trustworthy AI prompt cost estimates. We’re a small, operator-led team that publishes our assumptions, our sources, and our roadmap in plain sight.
Builder DNA
Why we care about tokens
We run real workloads for clients and internal products. Each time pricing moved or a new model landed, we rebuilt spreadsheets from scratch. TokenTally replaces that scramble with a single, auditable source.
Roadmap status
Foundation
Complete
Repo scaffolding, pricing plan, instrumentation, and builder-preview landing page.
Accuracy-first features
In progress
Pricing dataset, calculator engine, trust documentation, and public roadmap.
Scenario library
Up next
Deep-dive content, shareable presets, and practical guides for finance + engineering teams.
Operating principles
Accuracy over hype
Every estimate cites the underlying model pricing and assumptions so you can verify fast.
Explainable defaults
We publish the math behind each preset so finance, product, and engineering stay aligned.
Builder empathy
TokenTally is built by operators who run LLM workloads daily—we fix the pain we feel ourselves.
Why TokenTally exists
- No more inflating prompts with irrelevant context—we filter inputs locally before they ever touch an LLM.
- Finance leads need repeatable math; every scenario exposes its token counts and costs in plain English.
- Accuracy beats flash. When providers change pricing, we log the citation before updating the UI.