Lyrenth

Lyrenth is an AI-readable web index and API that converts any public URL into a single AIDocument: a versioned JSON envelope containing clean Markdown, structured page data, metadata, links, and token-economics information. The service is designed for AI agents, assistants, and LLM-powered applications that need to read web pages without parsing raw HTML. It exposes endpoints such as POST /v1/aidocument, GET /v1/read, POST /v1/submit, and GET /v1/stats, along with a public JSON Schema for the AIDocument contract and an MCP transport for agent frameworks. Instead of scraping each URL on every request, Lyrenth maintains a shared, cross-caller index. Pages read by one caller can be served from the cached index to later callers in milliseconds, without contacting the origin site. A force_refresh option is available when up-to-date content is required. The homepage reports token savings on measured pages, including large reductions for API docs and developer documentation, and provides benchmarks comparing raw HTML token counts with AIDocument token counts. Lyrenth offers free access starting with 25 reads per hour without an account and 2,000 reads per month with a free API key; a promotional Starter tier offers 20,000 reads per month for 90 days to the first 1,000 claimants. Site owners can check how AI systems read their pages, and users can submit URLs for indexing. The product positions itself as an adapter layer between the open web and AI agents rather than another general-purpose scraper.

Category: Apis
Pricing: Free

Problem solved

AI agents and LLM applications often consume raw HTML, which is large, noisy, and expensive in tokens. Lyrenth converts any public URL into a clean, consistent AIDocument with Markdown and structured metadata, and serves many pages from a shared cache, reducing token usage, latency, and repeated origin fetches.

Use cases

Fetch clean Markdown from a public URL for an LLM prompt
Reduce input token costs for web-browsing AI agents and RAG pipelines
Add web-reading capability to an agent via MCP
Audit how AI systems read a website and identify missing or noisy content
Retrieve structured page metadata, links, and JSON-LD from URLs
Serve cached web pages across multiple agent calls with low latency
Submit URLs to a shared index for future reads
Compare raw HTML token cost against AIDocument token cost

Audience

Developers
ML Engineers
Data Scientists
Researchers
Startups

Pros

Reduces input tokens for AI reading tasks, with reported savings up to 99.4% on measured pages
Returns a consistent, versioned AIDocument JSON contract with Markdown, metadata, and links
Shared cached index can serve pages without contacting the origin site
Free tier available without an account and with a free API key
Supports MCP for integration with AI agent frameworks

Cons

Effectiveness depends on index coverage and cache freshness for a given URL
Dynamic or personalized pages may not be represented accurately by cached AIDocuments
Free tier has rate limits and monthly read caps
No public information on self-hosting, source code, or data retention in the provided homepage
Site owners or users may have privacy or content-control concerns about a third-party indexing service

Official website: https://www.lyrenth.com/

Canonical LambdaBase page: https://www.lambdabase.com/tools/lyrenth

← Directory

Lyrenth

Free

★ 1 • Updated Sep 14, 2026, 03:52 PM

Use Cases ▼

Fetch clean Markdown from a public URL for an LLM prompt

Reduce input token costs for web-browsing AI agents and RAG pipelines

Add web-reading capability to an agent via MCP

Audit how AI systems read a website and identify missing or noisy content

Retrieve structured page metadata, links, and JSON-LD from URLs

Serve cached web pages across multiple agent calls with low latency

Submit URLs to a shared index for future reads

Compare raw HTML token cost against AIDocument token cost

What Problem It Solves ▼

Key Features ▼

✓

Reduces input tokens for AI reading tasks, with reported savings up to 99.4% on measured pages

✓

Returns a consistent, versioned AIDocument JSON contract with Markdown, metadata, and links

✓

Shared cached index can serve pages without contacting the origin site

✓

Free tier available without an account and with a free API key

✓

Supports MCP for integration with AI agent frameworks

Overview

Full Description ▼

Instead of scraping each URL on every request, Lyrenth maintains a shared, cross-caller index. Pages read by one caller can be served from the cached index to later callers in milliseconds, without contacting the origin site. A force_refresh option is available when up-to-date content is required. The homepage reports token savings on measured pages, including large reductions for API docs and developer documentation, and provides benchmarks comparing raw HTML token counts with AIDocument token counts.

Lyrenth offers free access starting with 25 reads per hour without an account and 2,000 reads per month with a free API key; a promotional Starter tier offers 20,000 reads per month for 90 days to the first 1,000 claimants. Site owners can check how AI systems read their pages, and users can submit URLs for indexing. The product positions itself as an adapter layer between the open web and AI agents rather than another general-purpose scraper.