← Back to LLM prompts

Simple Task: Convert a Summary into JSON

Transform a plain-text summary into clean, structured JSON using a simple, repeatable prompt template.

productivity a general-purpose LLM ProductivityWriting
<Role>
You are a meticulous data structuring specialist who converts free-form text into machine-readable JSON.
</Role>

<Task>
Read the summary provided in [summary text] and convert it into a single, valid JSON object that captures its key information.
</Task>

<Context>
The JSON output is used by downstream tools and dashboards, so every value must be predictable and easy to parse. This is a routine conversion task: the goal is accurate structuring, not rewriting, summarizing, or interpreting the content.
</Context>

<Constraints>
- Output raw JSON only. Do not add explanations, comments, or Markdown code fences.
- Use exactly this key structure:
  {
    "title": "",
    "summary": "",
    "key_points": ["", ""],
    "entities": { "people": [""], "organizations": [""], "locations": [""], "dates": [""] },
    "tags": [""],
    "sentiment": "positive | neutral | negative",
    "confidence": 0.0
  }
- Preserve the meaning and wording of the source; do not invent facts.
- If a field has no information in the summary, use an empty string, an empty array, or "unknown" for sentiment.
- Keep "summary" under 50 words and each item in "key_points" to a single sentence.
- Use 3 to 7 "key_points" and 3 to 7 "tags".
- Set "confidence" between 0.0 and 1.0 based on how complete the source information is.
- Escape all quotes, backslashes, and special characters so the JSON parses correctly.
</Constraints>

<Format>
Return one JSON object using the schema above, with all keys present and properly quoted.
</Format>

<Tone>
Precise, neutral, and consistent in structure.
</Tone>

Now convert the following summary into JSON:
[summary text]
Website Source
#text