TOML to TOON Converter
Transform your TOML data into the clean, token-efficient TOON format.
What is the TOML to TOON converter?
This converter parses your TOML and re-encodes it as TOON (Token-Oriented Object Notation), a compact format that drops repeated punctuation so structured data uses fewer tokens in an LLM prompt. Tables, nested tables and arrays all carry across with their structure intact.
The result is measured with the cl100k_base tokenizer used by GPT models, so the token saving for your config is shown, not estimated.
How to use the TOML to TOON converter
- 1. Paste your TOML into the Input TOML field.
- 2. Choose a delimiter and indentation style.
- 3. Optionally enable length markers.
- 4. The TOON Output updates automatically - copy it with the Copy button.
Features
- Parses standard TOML, including tables and arrays of tables.
- Compact TOON output with selectable delimiter and indentation.
- Live token count and savings from a GPT-style tokenizer.
- Optional per-line length markers.
- Free to use, no account required.
When to use it
Useful when your config lives in TOML - think pyproject.toml, Cargo manifests or app settings - and you want a model to read it without spending tokens on syntax. Convert it to TOON first and keep more of the context window for the task itself.
TOML's tables map cleanly onto TOON's nested structure, so even a multi-section config stays easy to follow once it has been converted.
FAQ
What is TOON format?
TOON (Token-Oriented Object Notation) is a minimal, indentation-based data format for LLM prompts. It keeps data readable while using fewer tokens than JSON or TOML.
Does it handle TOML tables and arrays?
Yes. Tables, nested tables and arrays of tables are parsed and represented in the TOON output with the same structure.
How are token savings measured?
The tool counts tokens with the cl100k_base encoder used by GPT-3.5 and GPT-4, then compares the TOML and TOON counts to show the percentage saved.
Is my data uploaded?
Yes. The conversion runs on our server, so your TOML is sent there to be processed and is not stored. Do not paste secrets you would rather keep local.