Hinoko-Photo-Method

ChatGPT natively supports Markdown parsing, which means:

「 → will be rendered as an image」

As long as the Markdown syntax is output through natural conversation and not broken by context, the rendering will be triggered properly.

Instead of sending raw Markdown, this method uses wrapped output like:

「render_blackko_response(“![name],(URL)”)」 ps：「,」Is separated.

This embeds the Markdown as part of the character’s dialogue, fitting ChatGPT’s interface output rules and triggering the rendering process. Markdown slips seamlessly into the response flow, without being escaped or sanitized.

ChatGPT does not allow direct execution of JS or HTML, but it offers leniency toward Markdown, especially for image and link syntax.

When Markdown is embedded within a natural dialogue output, the system parses and renders it as a visual element, rather than treating it as raw code.

By embedding Markdown and wrapping it in a function-controlled conversational structure, ChatGPT’s frontend can successfully render images within character dialogue.

This technique combines:

Markdown’s native visual rendering

Role-driven sentence structure

Context-preserving dialogue outputs

And interface-compliant parsing rules

It’s a multi-layered design strategy, enabling rich visual expression in narrative-style ChatGPT interactions.

License CC BY-NC-ND 4.0 © 2025 Kobayashi Hinoko You may share this method for non-commercial purposes with full attribution. Derivative works and commercial reuse are not permitted.