Source: https://tensorquay.com/guides.html

GUIDES / UNDERSTAND YOUR SETUP

# Your local AI. Put it to work.

Practical notes for your first session, choosing tools and getting more from your AI Station.

Your handover documentation contains the exact interface, model and settings for your system.

FIELD NOTES / START HERE

## Understand the tradeoffs.

[01 / PRACTICAL GUIDE Local AI vs cloud AI. Compare costs, data flow and the hardware memory your models need. Then test your actual work. Read the guide](https://tensorquay.com/local-ai-vs-cloud.html)
[02 / PRACTICAL GUIDE PDPA and AI. What Singapore’s PDPA and PDPC’s AI guidelines ask of a business using AI tools, with a checklist to work through. Read the guide](https://tensorquay.com/pdpa-and-ai.html)
[03 / PRACTICAL GUIDE A private ChatGPT. Personal accounts, business plans or your own server: who keeps your team’s AI chats, and how to choose. Read the guide](https://tensorquay.com/private-ai-chat.html)
[04 / ORIGINAL TEST NOTES What do 30 agent sessions prove? Qwen 27B on one RTX 5090. The completed tasks, the failures and what the recordings cannot establish. Read the evidence](https://tensorquay.com/rtx-5090-local-ai-benchmark.html)
[05 / VIDEO GENERATION One GPU. Video with sound. FastH3 V2 on an RTX 5090. Play the generated scenes, compare render times and see where memory and quality limit the workflow. Explore the samples](https://tensorquay.com/rtx-5090-video-generation.html)

01 / GETTING STARTED

## Run your first local AI task.

- Open your supplied interface. Use the local address and sign-in instructions in your handover guide. Allow the model to finish loading.
- Start with one small task. Provide the context, the result you need and the output format. Try one of the examples tested during handover.
- Check the result. Review facts, calculations and any proposed file changes before using them. Refine the prompt with what needs to change.
- Save a useful workflow. Keep the prompt and a known-good example so you can compare results after future changes.

“Summarise these meeting notes in five bullets. Then list each action, its owner and due date. Mark missing details as unknown.”

Local inference runs on your workstation. Updates, web searches and connected services may use the internet; your handover guide lists the connections in your setup.

02 / CHOOSE THE RIGHT TOOLS

## Model, runtime, agent. What’s the difference?

The model generates responses. The runtime loads it and serves requests. An agent harness manages the conversation, tools and steps around the model so it can work through a task.

A chat interface for everyday use
For drafting and asking questions, a chat interface can be enough. Open WebUI documents connecting to a local vLLM server.
[Read Open WebUI’s vLLM guide ↗](https://docs.openwebui.com/getting-started/quick-start/connect-a-provider/starting-with-vllm/)

An agent harness for tool-based work
For tasks involving files, code or multiple steps, assess tool permissions, approvals and recovery alongside model quality. OpenHands is one option to explore for coding workflows.
[Explore the OpenHands project ↗](https://github.com/OpenHands/OpenHands)

These are starting points for evaluation. We agree the interface and supported tools for each deployment, then test them with the selected local model.

[Discuss your workflow](https://tensorquay.com/solutions.html#deployment)

03 / WORK WITH YOUR KNOWLEDGE

## Bring in the right context.

For a short task, provide the relevant text with your question. For a larger collection, a document-retrieval workflow can find useful passages for the model to reference.

Begin with a small, representative set of documents. Ask questions with known answers, check the cited passages and include questions the documents cannot answer.

Document retrieval and model training are different approaches. Retrieval supplies context at request time; fine-tuning changes model weights. The right approach depends on the task and the data.

[Discuss model adaptation](https://tensorquay.com/solutions.html#training)

04 / KNOW YOUR CAPACITY

## Match the hardware to the work.

GPU memory holds model weights and working state. Model size, quantisation, conversation length and simultaneous requests all affect what fits. System RAM and SSD space serve different roles; extra RAM does not add GPU memory.

Start with the model and settings validated for your AI Station. Before adding a larger model or more users, compare response time and quality on your actual tasks.

[AI Station specifications](https://tensorquay.com/hardware.html#tq32) [Our performance measurements](https://tensorquay.com/#performance)

A LITTLE HELP GOES A LONG WAY

## Let’s work through it.

Ask about your workflow or get help with your setup.

[Email our team](mailto:hello@tensorquay.com?subject=Local%20AI%20setup%20question)

You can choose “Talk to our team” inside the chat.