ModelHarbor launched as a self-hosted model library on GitHub
Ernos Labs released a free, self-hosted archive of open model weights, allowing users to download and store models locally. The service aims to provide a reliable backup for open-weight models that may be removed from original sources.
Ernos Labs released a free, self-hosted archive of open model weights, allowing users to download and store models locally. The service aims to provide a reliable backup for open-weight models that may be removed from original sources.
The Hugging Bay is a website for downloading machine learning models as a backup in case Hugging Face censors or limits access. It mirrors model availability so users retain access to weights. It targets developers and researchers worried about platform-side restrictions.
The Hugging Bay is a website for downloading machine learning models as a backup in case Hugging Face censors or limits access. It mirrors model availability so users retain access to weights. It targets developers and researchers worried about platform-side restrictions.
A desktop app that acts as a command center for running multiple local LLMs from a single interface. It lets users manage and interact with local models without switching between separate tools, targeting developers and hobbyists who self-host models.
A desktop app that acts as a command center for running multiple local LLMs from a single interface. It lets users manage and interact with local models without switching between separate tools, targeting developers and hobbyists who self-host models.
Agenta is an open-source alternative to Claude Cowork that allows users to run self-hosted models and integrate any harness. It provides a flexible, self-hosted option for AI agent workflows.
Agenta is an open-source alternative to Claude Cowork that allows users to run self-hosted models and integrate any harness. It provides a flexible, self-hosted option for AI agent workflows.
Meta released open weights for Muse Glimmer-30B, an open-weight model anyone can download and run on their own hardware, contrasting with its more powerful Muse Spark, which remains API-only. The release accompanied a 6,500-word essay by Mark Zuckerberg arguing AI should be 'for everyone.' The model targets local agentic AI and is already available in Cline via Ollama.
Meta released open weights for Muse Glimmer-30B, an open-weight model anyone can download and run on their own hardware, contrasting with its more powerful Muse Spark, which remains API-only. The release accompanied a 6,500-word essay by Mark Zuckerberg arguing AI should be 'for everyone.' The model targets local agentic AI and is already available in Cline via Ollama.