JSON Mode — Animate | AI Wisdom

← Back to topic
🎞 ANIMATE · Mode 3 of 5 · +100 XP— JSON Mode
⚙️Configure
💬Prompt
📤Generate
✅Parse
🎯 Next: Test →
Step 1 / 4
Speed
Enable JSON mode
response_format: json_object forces valid JSON output💡 Analogy
Like asking for a receipt — you know it'll be in receipt format, but the items vary.
Setting response_format to json_object tells the model to always return valid JSON.
Console Output
> response_format: { type: "json_object" }
> Model will return: valid JSON
> Schema: not enforced
●json-mode.ts
Enable JSON modeL1–3
1const response = await openai.chat.completions.create({
2  model: 'gpt-4o',
3  response_format: { type: 'json_object' },
4  messages: [
5    {
6      role: 'system',
7      content: 'Return JSON only. Extract: name, age, skills array.',
8    },
9    {
10      role: 'user',
11      content: 'Alice, 30, knows Python, TypeScript and Docker.',
12    },
13  ],
14});
15 
16const data = JSON.parse(response.choices[0].message.content!);
17// { name: "Alice", age: 30, skills: ["Python", "TypeScript", "Docker"] }
18 
19console.log(data.skills.length); // 3
19 lines
Timeline — 4 Steps
Enable JSON mode
response_format: json_object forces valid JSON output💡 Analogy
Like asking for a receipt — you know it'll be in receipt format, but the items vary.
Setting response_format to json_object tells the model to always return valid JSON.
Console Output
> response_format: { type: "json_object" }
> Model will return: valid JSON
> Schema: not enforced

response_format: json_object forces valid JSON output

response_format: json_object forces valid JSON output

Was this page helpful?

Discussion

response_format: json_object forces valid JSON output

response_format: json_object forces valid JSON output

Was this page helpful?

Discussion