AI assistant for Hugging Face
Connect Hugging Face so Joshu can check model repositories, monitor inference spaces, and track machine learning datasets.
Then $95 a month with no token fees, no resets, and no caps.
What you can ask Joshu in Hugging Face
You: What model repositories did our team update on Hugging Face this month?
Joshu inspects your organization repositories, lists recently pushed model commits and README updates, and leaves a brief on your desk.
You: Summarize the splits and sample rows of our evaluation dataset.
Joshu reads the dataset info, splits, and first rows from Hugging Face, and leaves a short summary with row counts on your desk.
You: Check the deployment status and resource use of our Hugging Face Space.
Joshu reads the Space runtime status, recent status events, and resource metrics, and flags any startup errors or memory limits it finds.
Model weights, fine-tuned checkpoints, and evaluation datasets on Hugging Face connect directly to your engineering workflow.
What Joshu does
Joshu keeps track of your Hugging Face organization to monitor model card documentation, inspect dataset versions, and check the runtime status of inference Spaces. When new checkpoints are pushed, Joshu reads the commits and model card, flags Space failures it finds on its checks, and drafts release summaries for your engineering team.
What you do
Connect your Hugging Face account with OAuth and select which repositories, datasets, or Spaces Joshu may inspect. Decide whether Space and endpoint checks report into your daily brief or stay on call for specific queries. Joshu asks you to approve each commit, settings change, or post before it goes through.
Why it sticks
Monitoring open-source model repositories and Spaces across team members often requires checking multiple dashboards. Joshu inspects commit activity and container health directly so your machine learning assets stay organized.
Permissions and limits
Joshu accesses Hugging Face through the OAuth access you grant. It reads repository metadata, checks Space status and metrics, and runs chat completion calls within that access. It cannot commit to a repository or change its visibility without explicit confirmation.
Frequently asked questions
Joshu connects to Hugging Face through Composio-managed OAuth, accessing the model repositories, datasets, and Spaces you authorize. It checks deployment health, tracks model card updates, verifies dataset releases, and connects machine learning workflows back to your engineering tickets. You work with an assistant who has a desk, a computer, and persistent memory, not another machine learning hub to navigate for simple status checks.
Yes. Every Joshu runs in its own isolated, sandboxed cloud computer. Proprietary models, evaluation datasets, and logs Joshu reads stay in that private sandbox. Joshu only accesses repositories you authorize, and your data is never used to train public AI models.
No. Commits, Space settings and secrets, repository visibility changes, and discussion posts wait for your approval. This connection cannot delete a model repository. Joshu reads repositories, datasets, and Space status on its own.
Only the organizations, repositories, datasets, and Spaces your Hugging Face account can reach through the OAuth access you grant. Joshu cannot view private models or internal repositories in unconnected accounts.
Hugging Face tools operate within specific model hubs and Spaces. Joshu is an assistant with long-term memory across Hugging Face, GitHub, Slack, email, and task boards. It connects machine learning model releases to engineering documentation and checks your Spaces and endpoints on a schedule, then tells you when one reports an error.
Joshu is $95 a month after a 14-day free trial. That includes your sandboxed cloud computer, a dedicated US phone number, an @joshu.me email address, persistent memory across your tools, and all tool connectors, including Hugging Face. There are no token fees. Paid data and media APIs from the tool catalog are extra, paid per call from a prepaid usage wallet.