ThalesMMS

Scout / Support

Scout support

Setup, troubleshooting, and a direct contact for help.

Prepare your workspace

Select a report workspace for saved documents and configure corpus sources separately. The workspace, phrase index and structured-report definitions serve different purposes.

Save a new report to a file or configure its workspace before relying on autosave. Watch the save status for errors.

Set up completion

Import your shortcuts and corpus, then build the local phrase index. Configure embeddings separately when you want semantic retrieval.

AI completion can run after an idle pause following a word boundary when enabled and no local suggestion is active. Disable its toggle if you only want local suggestions.

Choose an AI provider

Use a configured OpenAI-compatible service or the Codex account flow. Check readiness and tool support before using an agent workflow.

The language-model provider does not replace the embeddings endpoint. Missing or failing embeddings need separate attention.

Troubleshoot dictation

For local dictation, configure a valid Whisper streaming executable and model and grant microphone access. The installer downloads required software and model files.

For remote dictation, check the speech endpoint and microphone permission. Refinement uses the selected language model even when speech recognition itself is local; turn refinement off if you want to avoid that text transfer.

Review generated text

Review every AI completion, dictated passage and structured-report proposal before using it. Scout is a clinical prototype and does not validate clinical findings or establish diagnostic correctness.

Contact support

Contact Thales Matheus M Santos about Scout at thalesmmsradio@gmail.com.

Include the app or package version, your operating system, the steps to reproduce the problem, and a sanitized error message. Do not send patient information, private documents or source code, API keys, passwords, or private keys.