How are tokens estimated?

With a fast character-based heuristic of roughly four characters per token, which is close enough to spot balance and outliers. For exact billing counts, use a real tokenizer for your model.

What message shapes are supported?

Each item must be an object with a role and a content field. Content can be a plain string or an array of content blocks (text parts are concatenated), covering both OpenAI and Anthropic styles.

Why does role balance matter?

A lopsided ratio — for example many assistant turns with little user input — can signal a runaway loop or a prompt that is not eliciting real interaction. It is a quick health check on a transcript.

Why flag long individual turns?

One enormous turn can dominate the context window and dilute the model's attention to the rest of the conversation. Flagging outliers helps you decide what to summarize or drop.

Conversation Turn Counter & Analyzer

Analyze a conversation’s structure and token balance

The Conversation Turn Counter & Analyzer takes a chat messages array and breaks down its shape: how many turns, the split between user and assistant, approximate tokens per turn, and which individual turns are unusually long. It is a quick health check before you send a transcript to a model or store it as training data.

How it works

The tool parses your JSON and reads each message’s role and content. String content is used directly; array content (Anthropic-style content blocks) has its text parts concatenated. Tokens are estimated with a character-based heuristic of roughly four characters per token — fast and good enough to compare turns and catch outliers. It then computes totals, the user/assistant ratio, the average tokens per turn, and flags any turn far above that average.

Tips and notes

Use this to sanity-check transcripts before fine-tuning or evaluation: a healthy dialogue usually alternates roles with reasonably balanced lengths. A single giant turn is a candidate for summarization so it does not crowd the context window on the next call. The token figures are estimates for structural insight, not billing — pair this with the OpenAI or Anthropic usage parsers when you need exact counts. Everything runs locally; nothing you paste leaves the browser.