A public map of the AI ecosystem.
Artificial Index is an independent encyclopedia and structured registry for understanding AI systems across products, models, organizations, APIs, research, capabilities, releases and historical lineage. It makes a fast-changing field easier to explore without pretending that every fact is equally complete or equally certain.
What the index is built to answer
What is this system? Who develops it? Is it active, historical or discontinued? How can people access it? Does it expose an API? Which modalities, integrations, deployment options and governance information are publicly documented? What evidence supports each answer?
The catalog provides broad coverage. Deep profiles add structured, field-level research so visitors can distinguish a known fact from an incomplete lead or an unanswered question.
What counts as an indexed AI
The registry includes named public products, foundation and specialist models, developer platforms, research systems, infrastructure, historically important projects and notable versions when the distinction is useful.
The total number of AI systems ever created is unknowable. The catalog count describes Artificial Index coverage, not a claim that a central authority has counted every AI in existence.
Evidence before certainty
Verified
The field is supported by an appropriate source and has passed the project’s review standard. Official documentation is preferred for current product, access, API, pricing and policy claims.
Partial
Useful evidence exists, but it does not fully answer the field, may be indirect, or needs further confirmation before the claim can be treated as complete.
Pending
The field has not yet been adequately researched or sourced. A pending field stays visible instead of being filled with an unsupported guess.
Status and access
Status describes the system’s current public availability or release state, such as public, preview, discontinued or historical.
Access describes the channels and conditions for using it, such as web, app, API, local download, waitlist, paid plan or enterprise agreement.
These fields are separate from uptime, deployment, licensing and lifecycle history.
What we do not publish
- Leaked credentials, private keys or confidential material.
- Unsupported claims presented as verified facts.
- Automated discoveries promoted without review.
- Completeness percentages inflated to make the catalog look finished.
- Product rankings disguised as neutral evidence coverage.
How updates enter the system
Automated discovery can identify releases, new companies and possible changes. AI can help classify and draft candidate records. Publication remains evidence-gated: candidates enter a review queue, sources are checked, conflicts are recorded and only supported fields are promoted.