Dimension scores are derived from public data and fields; weighted into the composite. Reference only.
CivArchive (formerly CivitAI Archive) is a cross-platform AI model search and metadata aggregation tool. It does not host model files itself. Instead, it crawls and indexes sources such as CivitAI, HuggingFace, TensorArt, SeaArt, PixAI, Shakker, TensorHub, Yodayo, Moescape, torrents, and others, bringing model information and download locations together on a unified page. It emerged because some models have been removed from centralized platforms like CivitAI, creating a community need for a more resilient search and archival directory.
The platform’s key technical feature is SHA256 hash matching: it uses the file hash as the unique ID rather than relying on names or platform links. This makes it possible to verify whether files from different sources are the same model and to consolidate multiple mirrors under a single record. Users can search by model name, keyword, tag, or SHA256, and can filter by model type, base model, tags, creator, and more. A unified model page lists known sources such as CivitAI, HuggingFace, and torrents, and shows whether the links are still accessible.
The crawled content does not reveal any pricing plan, membership tier, or free quota, so its business model cannot be determined. No public API documentation was found either. However, it provides stable /sha256/XXX links and is compatible with the original CivitAI URL paths. Contribution options include manually uploading to a HuggingFace repository or using a HuggingFace Space to automatically back up models from CivitAI.
Its main advantage is its clear positioning: as a “model directory” rather than a file repository, it reduces reliance on any single platform. The SHA256-based mechanism is useful for developers and researchers who need to verify model authenticity. Multi-mirror indexing and link health tracking can also help creators find alternatives when links go dead. The limitations are also clear: it does not store the model files themselves, so download availability depends entirely on third-party sources; sources such as HuggingFace may also remove content; account features, collections, recommendations, community verification, and metadata recovery are still under development; and information on privacy policies, support, and content governance is limited.
CivArchive is suitable for AI image/video creators looking for specific LoRAs, Checkpoints, or Workflows, as well as developers who need to verify model files by hash. The main text does not specify access conditions from China, and many of its external links depend on services such as HuggingFace, CivitAI, and torrents, so real-world usability may vary depending on the network environment. If it is unavailable, alternatives include using platforms such as CivitAI, HuggingFace, TensorArt, SeaArt, and PixAI directly, though each of these may also face content removals or access restrictions.
⚠ 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 civarchive.com official site.
civarchive.com is an Unknown Site Builders provider. TG4G tracks its product information, an overall rating of 8.0/10, and a China-accessibility score of China direct-connect friendly. Click "Visit Official Site" to reach civarchive.com directly.