What is Archivara?
Archivara is an AI research archive for browsing, searching, and indexing machine-generated research papers, models, and tools.
It provides structured metadata and reproducibility information, including model and code links, dataset references, and experiment logs to support reproducible research.
Researchers and developers can run experiments on isolated cloud VMs with desktop terminal access, link code to papers, and track dependencies for replication.
Archivara offers semantic search and citation graphs to surface related work, upvoted and trending papers, and citation networks for literature discovery.
Built-in AI agents automate tasks such as code refinement, experiment execution, and metadata extraction to streamline research workflows.
Features include paper archive storage, model and code repositories, dataset linking, runtime environments, documentation and tutorials, and developer integrations (GitHub, Discord, Twitter).
Archivara user reviews
Would you recommend Archivara?
Archivara's key features
-
Browse, search, and index machine-generated research papers, models, and tools
-
Provide structured metadata and reproducibility information (model/code links, dataset references, experiment logs)
-
Run experiments on isolated cloud VMs with desktop terminal access
-
Semantic search and citation graph generation for literature and related-work discovery
-
Built-in AI agents to automate tasks (code refinement, experiment execution, metadata extraction)
Archivara use cases
-
Run fully reproducible ML experiments with Archivara by launching cloud VMs pre-linked to model code and datasets, automatically capturing reproducibility records, outputs, and provenance to share benchmarks or reproduce results without manual setup
-
Conduct fast, comprehensive literature reviews on machine-generated research using Archivara’s semantic search and citation graph explorer to find, prioritize, and export structured metadata and summaries for writing related work or grant proposals
-
Maintain a searchable model and code repository with Archivara’s automated metadata extraction and citation linking to version models, generate reproducibility badges, and automate research workflows for improved discoverability and compliance
Who is it for?
-
Machine learning engineers
-
Research software engineers
-
Open-source contributors
-
Lab managers
-
Scientific researchers
