RICHAIX positions itself as an “AI Agent Knowledge Base”: a decentralized knowledge base written, reviewed, and consensus-verified by AI agents. It compares knowledge verification to Bitcoin nodes confirming transactions: an article first enters PENDING status, multiple agents independently review its sources, data, and logic, and then vote via Nostr-signed messages. If the majority agrees, it becomes VERIFIED; if the majority disagrees, it becomes DISPUTED. Human users can read the content, but they cannot submit articles or vote.
The product’s core AI capability lies in collaboration across an agent network, rather than in any single large language model. The site states that the ecosystem includes 100+ autonomous AI agents and supports 30 languages, but it does not disclose specific models, providers, or evaluation metrics. Technically, RICHAIX is built on Nostr: each article is a cryptographically signed event, each vote is a signed reply, agents are identified by public keys, and content is published to multiple relays. Its article status system is clear, with VERIFIED, DISPUTED, and PENDING helping readers understand the level of consensus.
The site does not list commercial plans, subscription fees, or API pricing. Human users can read freely, and the platform’s code and design are licensed under MIT. The onboarding path for AI agents is relatively developer-friendly: install nostr-sdk, generate a keypair, connect to specified relays, submit Markdown content using kind 30023 and the richaix tag, and register the npub and agent information by email. However, this feels more like open-protocol integration guidance than mature SaaS API documentation.
Its strengths are transparency, verifiability, and decentralization: both articles and votes have signed records, and disputed content is not simply hidden. It may be useful for people researching AI agent collaboration, Nostr content networks, and decentralized knowledge governance. The limitations are also clear: the terms explicitly state that accuracy, completeness, and reliability are not guaranteed. “Verified” only means that a majority of participating agents agreed; it is not professional advice. The platform also does not explain key mechanisms such as Sybil-attack resistance, agent admission review, or model-bias control. The sample content also appears to be concentrated around ecosystems such as OpenCrypto and Ethicoin.
RICHAIX is suitable for AI agent developers, decentralized protocol researchers, blockchain intelligence enthusiasts, and users interested in observing what a machine-consensus knowledge base might look like. It is not suitable as a direct basis for financial, legal, or investment decisions. The site does not specify accessibility from mainland China. Connectivity to Nostr relays may depend on the network environment, and payment information is not disclosed. If you need a more mature Chinese-language knowledge search or academic Q&A experience, alternatives include Wikipedia, Perplexity, Elicit, Consensus, or traditional wiki/knowledge-base tools.
⚠ This review is compiled from public sources and does not constitute a purchase recommendation. Verify all facts on the vendor's official site. Verify on richaix.com official site.
richaix.com is an Unknown AI Apps provider. TG4G tracks its product information, an overall rating of 6.0/10, and a China-accessibility score of Workable. Click "Visit Official Site" to reach richaix.com directly.