JSON extraction

Turn text into JSON at scalewith Qwen3.8 Flash 27B

Reviews, tickets, emails, and invoices come back as JSON fields. Check the keys before you use the object.

$1 after email verification. No card required.

Copy a request

curl and Python, ready to paste.

Both examples call https://www.flonno.com/v1 with model qwen3.8-27b, JSON mode, and reasoning off. Set FLONNO_API_KEY in the server environment. Check the keys before you use the object.

cURL
curl 'https://www.flonno.com/v1/chat/completions' \
  -H "Authorization: Bearer $FLONNO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "qwen3.8-27b",
  "messages": [
    {
      "role": "user",
      "content": "Extract a JSON object with the keys product, rating, sentiment, and issues from this product review: \"The kettle boils fast, but the lid sticks and the handle gets hot. I would still buy it again.\""
    }
  ],
  "max_tokens": 256,
  "reasoning_effort": "none",
  "response_format": {
    "type": "json_object"
  }
}'
Python
import os
from openai import OpenAI
 
client = OpenAI(
    api_key=os.environ["FLONNO_API_KEY"],
    base_url="https://www.flonno.com/v1",
    max_retries=0,
    timeout=120.0,
)
 
response = client.chat.completions.create(
    model="qwen3.8-27b",
    messages=[{
        "role": "user",
        "content": "Extract a JSON object with the keys product, rating, sentiment, and issues from this product review: \"The kettle boils fast, but the lid sticks and the handle gets hot. I would still buy it again.\"",
    }],
    max_tokens=256,
    response_format={"type": "json_object"},
    extra_body={"reasoning_effort": "none"},
)
print(response.choices[0].message.content)

One measured example

A product review, returned as JSON.

On 9 November 2026 one product-review extraction used 178 input tokens and 57 output tokens and finished in 1.6 seconds.

Published input rate$0.05
Published output rate$0.80
This measured call$0.0000545
100,000 calls at this size$5.45

One measured example. Figures apply $0.05 input and $0.80 output per million tokens to that token count. Taxes, retries and other workflow costs are separate.

Monthly cost

Price a day of calls.

Enter calls per day and the average input and output tokens. The 30-day total uses the published rate in the browser.

30-day cost$16.35

$0.05 input and $0.80 output per million tokens. Input and output start from the measured example. Change them for your own calls.

Before a large run

Context, storage, and the signup credit.

131,072 tokens

Input and output share one context window.

No prompt or completion storage

Request content is not written to content logs, disk caches, or training data.

$1 after email verification

No card for that credit. Open the dashboard after you verify your email.

Start with $1 on us.

Get $1 free