FAQ

Setup and Environment

Do I need internet access after initial setup?

You need internet access only to install dependencies and pull Ollama models. After that, SafeQueryAI can operate locally.

What software is required?

• .NET SDK 8.0+
• Node.js 18+
• Ollama

Which URLs should I expect in local development?

• Frontend: http://localhost:5173
• Backend API: http://localhost:5000/api
• Swagger: http://localhost:5000/swagger
• Ollama: http://localhost:11434

Product Behavior

What is SafeQueryAI used for?

SafeQueryAI supports document question-answering on uploaded PDF and CSV files using session-based processing.

Which file types are supported?

• .pdf
• .csv

What is the upload limit?

Configured upload limit is 20 MB per file, with a 25 MB request ceiling.

How long does a session last?

Sessions expire after 60 minutes of inactivity by default.

Can a user clear data before timeout?

Yes. The session clear action removes temporary files and in-memory index data immediately.

Security and Privacy

Where is document data processed?

Document processing occurs locally in the running environment through the backend and local LLM runtime.

Is document content sent to cloud services?

No cloud upload flow is implemented in the current architecture.

Is document data stored permanently?

No. SafeQueryAI uses temporary storage and in-memory session data that are removed on clear or expiry.

Is telemetry enabled?

No external telemetry integration is documented in the current implementation.

Does the app enforce local model inference?

Yes. Backend startup validates that the Ollama URL is a loopback address.

API and Development

Where can I find endpoint details?

See API Reference.

Is there a streaming endpoint?

Yes. Use POST /api/sessions/{sessionId}/questions/stream with text/event-stream.

Are there tests?

Yes. Backend and frontend test suites are documented in Testing.

Limits and Known Constraints

Does SafeQueryAI support OCR for scanned PDFs?

Not in the current release.

Is authentication built in?

Not in the current release.

Is this documented as a multi-tenant production platform?

No. Current documentation describes a local, privacy-first architecture with session-based processing.

Roadmap Questions

Are cloud deployment templates included today?

Not in the current release scope.

What improvements are planned next?

See Roadmap for planned reliability, retrieval, and operational improvements.

Reviewer Notes

• [REVIEW REQUIRED: verify whether any local environment uses a non-default backend port]