Developer tool / 32
System Prompt & Tool-Use Schema Builder
Build structured LLM system prompts with XML sections and formatted OpenAI/Claude tool-calling JSON schemas.
System Prompt & Tool-Use Schema Builder: TOEA formats system prompts into standard XML boundary tags (<role>, <instructions>, <constraints>) and generates OpenAI function calling specs and Anthropic Claude tool schemas. Runs 100% locally in your browser with zero server file uploads.
- Category
- Developer tools
- Runs
- In your browser
- Cost
- Free · no sign-up
- Availability
- Ready to use
Prompt Structure & System Sections
Tool-Calling Definitions
Descriptions decide which tool is called
A model chooses a tool from its name and description alone, so the description does most of the work. Say what the tool does, when to use it, and when not to, especially if two tools overlap. Parameter descriptions should give the expected format and units, with an example: ISO 8601 date, e.g. 2025-06-30 or amount in cents. Mark a parameter required only when the call cannot work without it; otherwise the model invents values to fill it.
Finishing the schemas by hand
The builder writes each parameter as a type and a description. An array parameter has no items and an object parameter has no properties, so the model is not told what goes inside; add those, plus enum for fixed choices. To build a nested parameter schema from an example payload, start from JSON to JSON Schema and paste its object in. Both formats wrap the same JSON Schema: OpenAI nests it under function.parameters, Claude calls it input_schema, so edits carry over between the two.
How to use it
- Fill in prompt sections (Role, Context, Instructions, Constraints, Output Format, Examples).
- Define optional tool-calling functions with parameters.
- Switch between XML Prompt, Markdown, OpenAI Tool Schema, and Claude Tool Schema tabs, then copy your formatted prompt.
Privacy & limitations
All prompt editing and schema generation runs locally in client memory.
Related tools
Frequently asked questions
Why use XML tags in system prompts?
Modern LLMs (Claude 3.5 Sonnet, GPT-4o) parse XML tags (<role>, <instructions>) effectively, reducing instruction drift and ambiguity.
Does it support both OpenAI and Claude tool formats?
Yes. It generates valid OpenAI function schemas and Claude input_schema definitions simultaneously.
Free tool · runs in your browser · no account required