OpenRouter
Routes API requests to 500 language models across 80 providers through a unified interface.
OpenRouter is an American artificial intelligence company that operates a platform for accessing and routing requests to large language models and other generative artificial intelligence models.First sentence of the English Wikipedia article
91/100
- Entity25/25
- Machine-readable31.5/35
- Retrieval access17/20
- Footprint17.2/20
- OpenRouter91
- LLM operations medianof 35 rated62
- Atlas medianof 537 rated61
- Retrieval botspartialverify
Whether the bots that decide citations are let in. Training bots are a separate list and are not counted here.
- llms.txtyesverify
A plain-text index at /llms.txt. Optional; read by some tools.
- Organization schemayes
schema.org Organization markup on the homepage, which is what a resolver reads a company's name and links from.
- Wikidata itemQ140795851Wikipedia article
Matched on the official website, never on the name. The substrate under the knowledge graph an engine resolves names against.
- Server-rendered words465swept
What a reader without JavaScript receives from the homepage.
- Identity score90/100swept
The analyzer's crawl-only section, 0-100: identity files, schema, crawler policy, sitemap. Not the analyzer score.
- Footprint2.8k GitHub stars
- HeadquartersNew York CityP159
90/100
identity score
Not measured with the analyzer.
The full run asks four engines about this company by name. It runs on request, not on a schedule.
What is OpenRouter?
Routes API requests to 500 language models across 80 providers through a unified interface.
What is OpenRouter's Atlas Readiness?
91 of 100 on 2026-09-15, computed from four parts — entity, machine-readable, retrieval access, footprint — all four parts read. It places #1 of 35 rated LLM operations tools.
Does OpenRouter let AI retrieval bots in?
Partly. On 2026-09-15, openrouter.ai fenced some paths from retrieval crawlers.
Does OpenRouter publish an llms.txt?
Yes: openrouter.ai/llms.txt was present on 2026-09-15.
Is OpenRouter a Wikidata entity?
Yes: Q140795851, matched on its official website.
How to read a row
- swept
- Read by a crawler: HTTP and parsing, no model call. Every row, every week.
- measured
- The full analyzer run, asked of four engines. Shown only where the company agreed to show it.
Can a retrieval engine find, resolve and quote this site? Four parts, each read by the sweep, each shown beside the total.
Entity 25 · Machine-readable 35 · Retrieval access 20 · Footprint 20
A part that was never read is left out and the total is rescaled to what was. It is never counted as zero.