AI Prep Tools
Token & Context Calculator
Check context fit, split prompts into chunks, and see which token counts are exact vs. estimated.
- Paste text or drop a text/JSON/Markdown file and choose a target model.
- Reserve your expected output tokens to see whether the input fits comfortably, runs tight, or overflows.
- Use the Chunk Splitter to break long inputs along paragraph and heading boundaries.
Loading local browser workspace...
Supported Formats, Limits & Accuracy
OpenAI models using local o200k_base / cl100k_base BPE run Exact in-browser; Claude and Gemini rows show Estimated ranges with documented error bands.
- Supported:
- Pasted text, .txt, .md, .json, .csv
- Processing:
- Client-side browser execution ( how to verify)
Honest Limitations
- Anthropic Claude and Google Gemini tokenizers are not bundled as open BPE tables; counts for those models are honest estimates rather than API queries.
- Code, JSON, and non-Latin scripts vary much more across tokenizers than plain English prose.
How to Use Token & Context Calculator
- Paste text or drop a text/JSON/Markdown file and choose a target model.
- Reserve your expected output tokens to see whether the input fits comfortably, runs tight, or overflows.
- Use the Chunk Splitter to break long inputs along paragraph and heading boundaries.
Frequently Asked Questions
- Why doesn't this tool call vendor APIs for Claude or Gemini counts?
- Sending your text to a remote API just to count tokens breaks our zero-upload guarantee. Instead we run open tokenizers locally for Exact OpenAI counts and label proprietary counts clearly as Estimated.
Next Steps & Related Tools
Core Tool
Open tool Word & Character Counter
Live word, character, sentence, paragraph, and reading-time stats with transparent counting rules.
AI Prep
Open tool Safe-to-Paste Redactor
Mask emails, phones, cards, IBANs, IPs, and API keys locally — then restore originals in AI replies.
AI Prep
Open tool LLM JSON Repair & Schema Check
Repair malformed model JSON, inspect every change with per-rule undo, and validate schemas.