Set up Scrublet as a remote tool

Sharing this model for a user, not developing it? Follow the tooluniverse-share-computer guide; no ToolUniverse source checkout is needed. The model runs in its own Python environment, with its dependencies:

~/.local/bin/uv venv --python 3.12 ~/.tooluniverse/envs/scrublet
curl -fsSL -o ~/.tooluniverse/envs/scrublet-tooluniverse.txt https://connect.aiscientist.tools/install-requirements.txt
curl -fsSL -o ~/.tooluniverse/envs/scrublet-requirements.txt https://raw.githubusercontent.com/mims-harvard/ToolUniverse/main/src/tooluniverse/remote/scrublet/requirements.txt
~/.local/bin/uv pip install --python ~/.tooluniverse/envs/scrublet/bin/python -r ~/.tooluniverse/envs/scrublet-tooluniverse.txt -r ~/.tooluniverse/envs/scrublet-requirements.txt
~/.local/bin/tu remote check scrublet --python ~/.tooluniverse/envs/scrublet/bin/python   # lists anything still missing
~/.local/bin/tu serve --allow scrublet --python ~/.tooluniverse/envs/scrublet/bin/python --share --detach --name "<computer name>" --service https://tooluniverse-backend.onrender.com

If the check reports a missing environment variable (a data folder, a model cache, a provider key), set it in the same shell before tu serve; "Obtain credentials, data, and model weights" below says what each one is. The model reads only the variables it declares.

The rest of this guide is for developers working from the ToolUniverse source, and for diagnosing problems with this model.

Validation status (2026-08-16): installation, direct/live MCP calls, traversal rejection, and prior same-host concurrency levels 1, 2, and 4 passed. Two current 500-cell calls returned finite aligned results and 429 synthetic predictions; that 85.8% fixture rate is not an accuracy result. Public publication, cross-user isolation, representative accuracy, cancellation, and recovery remain incomplete. Authenticated private Platform import and owner testing passed on 2026-08-16; public publication and independent-caller authorization/isolation remain untested.