{}

Prompt Workbench

JSON Prompt Builder

Back to index

Templates

Humanization examples for social media JSON prompts

Examples showing how the humanization block changes the tone of a social post prompt and which AI markers are worth excluding.

Published: April 2, 2026Updated: April 2, 2026Reading time: 5 minAuthor: Astraport Team
json
{
  "task": "Write an Instagram post about the benefits of meditation",
  "language": "ru",
  "tone": "friendly",
  "output_format": "plain_text",
  "humanization": {
    "avoid_long_dashes": true,
    "allow_minor_imperfections": false,
    "avoid_ai_phrases": true,
    "markers": ["avoid_todays_world", "avoid_in_conclusion"],
    "custom_excluded_phrases": "в современном мире\nважно отметить"
  }
}

The humanization block is useful when a generated post sounds too polished, too generic, or too obviously AI-written.

What this parameter can control

  • Avoid long dashes
  • Exclude AI marker phrases
  • Allow minor imperfections when the channel is intentionally casual

Recommended baseline

Start with avoid_long_dashes = true and avoid_ai_phrases = true. Enable imperfections only when the brand tone already uses informal language.

json
{
  "task": "Write an Instagram post about the benefits of meditation",
  "language": "ru",
  "tone": "friendly",
  "output_format": "plain_text",
  "humanization": {
    "avoid_long_dashes": true,
    "allow_minor_imperfections": false,
    "avoid_ai_phrases": true,
    "markers": ["avoid_todays_world", "avoid_in_conclusion"],
    "custom_excluded_phrases": "в современном мире\nважно отметить"
  }
}
Open the SMM builder

FAQ

FAQ

Should humanization always allow mistakes?

No. Usually it is enough to remove AI markers and over-polished phrasing.

Which phrases are worth excluding first?

Start with broad patterns like “in today's world”, “it is worth noting”, and “in conclusion”.