PromptSpend

Every cost comparison tool becomes obsolete the moment vendors adjust their pricing. PromptSpend addresses this fundamental problem by treating the pricing pipeline itself as the core product rather than treating price data as static information to collect once.

The calculator supports 78 LLM models including GPT and Claude, allowing side-by-side comparison of up to four models at a time. Users can input real prompts or specify their usage scale to see monthly cost projections. The platform targets developers and technical teams making model selection decisions.

What distinguishes PromptSpend from existing calculators is its automated approach to freshness. The system re-checks pricing information every morning against vendor pages, flagging discrepancies instead of simply averaging different sources. Each price quote includes the source URL and confirmation date, providing verifiable attribution that existing calculators typically omit.

The tool also addresses two cost factors that most calculators ignore or get wrong. The first is compounding chat history—each message in a conversation includes previous messages, yet calculators often treat input costs as linear. The second is hidden reasoning tokens, which affect models that perform internal reasoning steps before generating output. These technical details significantly impact real-world costs at scale.

PromptSpend offers three access points beyond the web calculator. A public API lets developers integrate pricing lookups programmatically. An MCP server makes the tool available to coding agents, enabling AI assistants to check costs while writing code. A VS Code extension shows pricing information inline on the line of code that calls a model.

The pricing model is transparent: the tool is free, requires no account to use, and operates under an open source license. The founder chose this model deliberately, framing the continuously-updated pricing data as the key value proposition. Rather than monetizing access to price information or user data, the focus remains on the open source infrastructure itself being valuable to the ecosystem.

For teams running multiple LLM models in production or evaluating model choices, the combination of automated pricing accuracy and technical cost details addresses real friction in AI infrastructure decisions.