Schematron V2 Small
Schematron V2 Small is a 3B-parameter model developed by Inference.net designed specifically for extracting data from HTML into JSON format. It is optimized to handle complex schemas and long web pages, providing structured output based on user-supplied instructions.
Schematron V2 Small specs and pricing
- Provider: Inference.net
- Released: September 12, 2026
- Context window: 128K tokens
- Max output: 4K tokens
- Input: text
- Output: text
- Price: $0.05 input / $0.23 output per 1M tokens
What Schematron V2 Small is good at
- Converting large HTML documents into structured JSON data
- Extracting specific information from complex web pages
- Automating data parsing tasks based on defined schemas
To-dos to try with Schematron V2 Small
- Extract all product prices and availability statuses from this HTML page into a JSON format based on my provided schema.
- Parse the table data from this long HTML document and map it to the specified JSON structure.
- Identify and extract all contact information from this website's HTML source code according to my JSON template.
FAQ
- What is the primary function of Schematron V2 Small? This model is specialized in extracting information from HTML content and converting it into JSON format.
- Does this model support image inputs? No, Schematron V2 Small only accepts text as input.
- How much does Schematron V2 Small cost? Schematron V2 Small costs $0.05 per million input tokens and $0.23 per million output tokens through OpenRouter. In todo.is a typical agent step costs about 0.17 credits ($1 = 100 credits).
- What is the context window of Schematron V2 Small? Schematron V2 Small can read up to 128K tokens at once and write up to 4K tokens in one reply.
- Can I use Schematron V2 Small without coding? Yes. In todo.is you pick Schematron V2 Small as your agent's model and write to-dos in plain words, on the web, the Android app, WhatsApp or Telegram. No API key needed: credits are included, or you can bring your own key.
JavaScript is required to use the todo.is app.