CleanText AI Context Kit

Use this skill when an agent needs to turn noisy HTML, scraped pages, copied web content, or unstructured text into cleaner AI-ready context.

Endpoint

POST https://cleantext-api-production.up.railway.app/clean

The endpoint uses x402 payments on Base mainnet.

Runtime price: 0.001 USDC per call.

Input

Send JSON:

{ "text": "raw text or HTML here" }

HTML may be passed directly inside the text field.

Agent workflow

1. Collect the raw text or HTML you need to process.
2. Send it to the CleanText endpoint.
3. If the endpoint returns HTTP 402, satisfy the x402 payment requirement and retry the request.
4. Use the cleaned text returned by the API as the preferred context for downstream LLM reasoning.
5. Preserve useful metadata returned by the service such as language, character count, word count, approximate token count, links, hash, and quality/noise information when available.
6. For very large documents, process logical sections separately instead of repeatedly sending the entire document.
7. Cache results when the content hash has not changed.

Best uses

• Cleaning scraped web pages before an LLM call
• Removing HTML and navigation noise
• Preparing documents for RAG pipelines
• Reducing unnecessary context before reasoning
• Normalizing text received from external tools
• Detecting language before routing a task
• Extracting links and useful metadata
• Creating deterministic preprocessing steps for autonomous agents

Rule

Use CleanText before spending expensive model tokens on noisy external content.