picking-a-format

The best document format for LLM context depends on the consumer system. This skill maps consumer needs to the optimal text output and structured metadata format, choosing between text, markdown, djot, html, and JSON for efficient prompt context embedding.

To choose the right extraction format, align the consumer type with the optimal output pair. This skill maps systems like LLMs, RAG pipelines, parsers, and web displays to the ideal --format and --content-format combinations for efficient document processing.

Structured metadata extraction is needed when preparing documents for AI context workflows. This skill determines the appropriate format for extracted documents by optimizing for text output and structured metadata to ensure efficient data presentation for downstream consumers.

Yes, markdown or djot can be used for chatbot prompt context. This skill helps select the ideal format for document extraction by mapping the chatbot consumer to the optimal format and content-format pair for efficient prompt context embedding.

Yes, this format selection approach works for parsers and web display. The skill provides consumer-specific pairing, aligning the output format with the specific needs of different systems including parsers, archives, LLMs, and web display interfaces.

JSON for document extraction formatting is optimized for structured metadata but may lack the text readability needed for LLM context. This skill helps navigate format limitations by mapping your specific consumer to the optimal text or JSON output pair.