Tool
count_tokens
1 published MCP server declares count_tokens. llm-cost-estimator has the highest trust score right now. Servers describe it as: Token count for one model. Args: text, model?. OpenAI is exact (tiktoken); others are flagged estimates. Few published servers use this exact name, so read each listing before you install. Open a server for its install config and the rest of its tools.
45Trust score 45 of 100, Fair
Token counting & multi-model LLM cost estimates: GPT-4o, Claude, Gemini, 25+. No API key.
Registry
FAQ
- What is the count_tokens MCP tool?
- count_tokens is a tool name MCP servers declare. A description they publish is: Token count for one model. Args: text, model?. OpenAI is exact (tiktoken); others are flagged estimates.
- Which MCP servers provide count_tokens?
- 1 published server declares count_tokens. Highest in the current trust order: llm-cost-estimator.
- How do I use count_tokens?
- Install a server that declares count_tokens, then call it from Claude, Cursor, or VS Code. Each server page has the config for that client and the other tools that server exposes.