A minimal API to make LLMs do what you actually need
Our api makes mainstream llms very simple to use. To get json output:
curl -X POST "https://verysimplellm.aitida.com/json" \
-H "Content-Type: application/json" \
-H "Authorization: your-model-provider-key" \
-d '{
"model": "anthropic/claude-sonnet-4-20250514",
"prompt": "Please analyze the following user review: \"This coffee maker is absolutely amazing! It brews perfect coffee every morning and the design looks great in my kitchen. The only minor issue is that it takes a bit longer to brew than expected, but the quality makes up for it. Highly recommend!\"",
"json_schema": {
"type": "object",
"properties": {
"title": {
"type": "string",
"description": "A brief title summarizing the review"
},
"sentiment": {
"type": "number",
"description": "Sentiment score from -1 (negative) to 1 (positive)"
},
"category": {
"type": "string",
"enum": ["electronics", "home", "food", "clothing", "books", "other"],
"description": "Product category"
}
},
"required": ["title", "sentiment", "category"],
"additionalProperties":false
}
}'const response = await fetch('https://verysimplellm.aitida.com/json', {
method: 'POST',
headers: {
'Content-Type': 'application/json',
'Authorization': 'your-model-providers-key'
},
body: JSON.stringify({
model: 'anthropic/claude-sonnet-4-20250514',
system_prompt: 'Only output required fields.',
prompt: 'Please analyze the following user review: "This coffee maker is absolutely amazing! It brews perfect coffee every morning and the design looks great in my kitchen. The only minor issue is that it takes a bit longer to brew than expected, but the quality makes up for it. Highly recommend!"',
json_schema: {
type: 'object',
properties: {
title: {
type: 'string',
description: 'A brief title summarizing the review'
},
sentiment: {
type: 'number',
description: 'Sentiment score from -1 (negative) to 1 (positive)'
},
category: {
type: 'string',
enum: ['electronics', 'home', 'food', 'clothing', 'books', 'other'],
description: 'Product category'
}
},
required: ['title', 'sentiment', 'category'],
additionalProperties:false
}
})
});
const data = await response.json();
console.log(data);import requests
response = requests.post('https://verysimplellm.aitida.com/json',
headers={
'Content-Type': 'application/json',
'Authorization': 'your-model-providers-key'
},
json={
'model': 'anthropic/claude-sonnet-4-20250514',
'system_prompt': 'Only output required fields.',
'prompt': 'Please analyze the following user review: "This coffee maker is absolutely amazing! It brews perfect coffee every morning and the design looks great in my kitchen. The only minor issue is that it takes a bit longer to brew than expected, but the quality makes up for it. Highly recommend!"',
'json_schema': {
'type': 'object',
'properties': {
'title': {
'type': 'string',
'description': 'A brief title summarizing the review'
},
'sentiment': {
'type': 'number',
'description': 'Sentiment score from -1 (negative) to 1 (positive)'
},
'category': {
'type': 'string',
'enum': ['electronics', 'home', 'food', 'clothing', 'books', 'other'],
'description': 'Product category'
}
},
'required': ['title', 'sentiment', 'category'],
'additionalProperties':False
}
}
)
print(response.json()){
"json": {
"title": "Excellent Coffee Maker with Minor Brewing Speed Issue",
"sentiment": 0.8,
"category": "home"
}
}Supports any model LiteLLM supports, such as OpenAI, Anthropic, Google and more. Grab model identifiers from https://docs.litellm.ai/docs/providers
Put your LLM providers Api Key into the Authorization header. We dont store this anywhere ever.
View the source code at https://github.com/joshp-f/verysimplellm