Why these exist
Most of what I write here comes from running agents on real work and writing down what it cost and what broke. These tools are the calculators and checklists I kept rebuilding in spreadsheets for that work, cleaned up so you can use them too.
Everything runs in your browser. Nothing you type is sent anywhere. Share links keep your inputs in the part of the URL after the #, which browsers do not send to the server.
Where to start
- Choosing a model? Start with the model finder, then check the shortlist in the pricing explorer.
- Writing a budget? The cost simulator shows why per-call estimates come in low: every step re-sends the context so far.
- Designing a system? Draw it in the architecture playground and read the design review before anyone writes code.
- About to ship? Take the readiness test with your team and compare answers.
Where the data comes from
Model prices, context windows and capability flags come from LiteLLM’s public model price list and are cross-checked against OpenRouter’s model API once a day. When the two disagree, the tools say so. Benchmark scores are the Artificial Analysis indexes that OpenRouter publishes. Before you commit a budget, check the provider’s own pricing page; every row links to it.
The raw, normalized dataset is public at /wp-json/dp-tools/v1/models if you would rather build your own view.
Do I need an account?
No. There is no sign-up, no email gate and no login. The tools work as soon as the page loads.
Is anything I enter stored?
Inputs stay in your browser. A few conveniences — your last pricing visit, a saved scenario, a playground draft — are kept in your browser’s local storage so they are there when you come back. Anonymous usage events (for example “calculation completed”) go to the site’s existing analytics, with no personal data.
How current are the prices?
The dataset is refreshed daily. Each tool shows the date the data was last checked, and the pricing explorer keeps a log of price changes it has seen.