README.md

l1m 🚀

A Proxy to extract structured data from text and images using LLMs.

🌟 Why l1m?

l1m is the easiest way to get structured data from unstructured text or images using LLMs. No prompt engineering, no chat history, just a simple API to extract structured JSON from text or images.

✨ Features

• 📋 Simple Schema-First Approach: Define your data structure in JSON Schema, get back exactly what you need
• 🎯 Zero Prompt Engineering: No need to craft complex prompts or chain multiple calls. Add context as JSON schema descriptions
• 🔄 Provider Flexibility: Bring your own provider, supports any OpenAI compatible or Anthropic provider and Anthropic models
• ⚡ Caching: Built-in caching, with x-cache-ttl (seconds) header to use l1m.io as a cache for your LLM requests
• 🔓 Open Source: Open-source, no vendor lock-in. Or use the hosted version with free-tier and high availability
• 🔒 Privacy First: We don't store your data, unless you use the x-cache-ttl header
• ⚡️ Works Locally: Use l1m locally with Ollama or any other OpenAI compatible model provider

🚀 Quick Start

Text Example

curl -X POST https://api.l1m.io/structured \
-H "Content-Type: application/json" \
-H "X-Provider-Url: demo" \
-H "X-Provider-Key: demo" \
-H "X-Provider-Model: demo" \
-d '{
  "input": "A particularly severe crisis in 1907 led Congress to enact the Federal Reserve Act in 1913",
  "schema": {
    "type": "object",
    "properties": {
      "items": {
        "type": "array",
        "items": {
          "type": "object",
          "properties": {
            "name": { "type": "string" },
            "price": { "type": "number" }
          }
        }
      }
    }
  }
}'

Image Example

curl -X POST https://api.l1m.io/structured \
-H "Content-Type: application/json" \
-H "X-Provider-Url: demo" \
-H "X-Provider-Key: demo" \
-H "X-Provider-Model: demo" \
-d '{
  "input": "'$(curl -s https://public.l1m.io/menu.jpg | base64)'",
  "schema": {
    "type": "object",
    "properties": {
      "items": {
        "type": "array",
        "items": {
          "type": "object",
          "properties": {
            "name": { "type": "string" },
            "price": { "type": "number" }
          }
        }
      }
    }
  }
}'

Node.js Example

See sdk-node for a complete example.

Python Example

See sdk-python for a complete example.

Go Example

See sdk-go for a complete example.

📚 Documentation

API Headers

• x-provider-model (optional): Custom LLM model to use
• x-provider-url (optional): Custom LLM provider URL (OpenAI compatible or Anthropic API)
• x-provider-key (optional): API key for custom LLM provider
• x-cache-ttl (optional): Cache TTL in seconds
  • Cache key (generated) = hash(input + schema + x-provider-key + x-provider-model)

Supported Image Formats

• image/jpeg
• image/png

🛠️ SDKs

Official SDKs are available for:

• Node.js
• Python
• Go

⚡️ Running Locally

See local.md for instructions on running l1m locally (and using with Ollama).

🔔 Stay Updated

Join our waitlist to get early access to the production release of our hosted version.

📚 Acknowledgements

• BAML
• Zod
• ts-rest
• ajv

🏢 About

Built by Inferable.