Getting Started
The best way to understand Alfrada OS is not as a chatbot, but as an operating system for work.
Most work does not live in one message. It lives across drafts, files, research, revisions, outputs, and next steps. Alfrada OS is built to keep that work together around you, the user.
That means sessions, files, and context can be pulled forward from earlier conversations, so the work does not have to restart from zero each time. A new conversation can feel less like a blank page and more like a continuation of where you last left off.
Start With The Outcome
Strong sessions usually begin with a clear outcome, not a vague topic.
Instead of asking for “thoughts on the market,” ask for a market brief. Instead of asking for “help with this file,” ask for a summary, a cleaning plan, or a presentation outline. The clearer the destination, the better the result.
In practice, one of the best first moves is simply to ask Alfrada OS whether it can help you achieve the objective you have in mind.
Because Alfrada OS is self-planning, it can often tell you the smartest path forward before you know which tool, mode, or workflow to ask for. In many cases it can also handle basic in-product actions on your behalf, including changing simple settings or helping you complete setup flows without sending you into a settings page first.
I want to achieve [objective]. Can you help me get there? Please tell me the best way to approach this in Alfrada OS, what context you need from me, and what the first step should be.Bring The Raw Material Early
If you already have a file, a draft, a spreadsheet, a website, or linked data, add it early.
Good source material improves the work immediately. It gives Alfrada OS something concrete to analyze, transform, or build on, which usually leads to better answers and fewer follow-up questions.
Let The Session Hold The Work
A session is more than a chat thread. It is the working container for the job itself.
As the session develops, your uploads, outputs, generated files, and follow-up requests stay connected. That means you can refine the work without repeatedly re-explaining the context. Finished artifacts belong in the side workspace, not buried in the chat scroll.
Continue From Earlier Work
One of the most useful habits is to tell Alfrada OS to look backward before it moves forward.
If something was discussed last week, if a file lived in an older thread, or if you only remember part of the name, ask for it directly. Conversation history is part of the working context.
Search my conversation history and find `syz.docx`. I think this was about 5 days ago. Pull the relevant context into this session and tell me what it was about.Search my conversation history for the last time we worked on [project or client]. Summarize where we left off, what files mattered, and what the smartest next step is.Look through my previous conversations for the draft we made about [topic]. Bring back the most relevant version, summarize it briefly, and help me continue from there.Learn The Interface As You Need It
You do not need to master every feature on day one.
The simplest path is to start with one task, let Alfrada OS propose the path, review the output, and iterate. As you go, you will naturally use more of the workspace: files, drafts, generated outputs, connected data sources, playbooks, and automations. And when a job needs a follow-up rather than a schedule, Alfrada OS can wake the same conversation back up later to check on it — see Automations.
Playbooks Make The System More Useful Over Time
Alfrada OS does not start from a blank workflow library.
The product includes substantial playbook coverage for real jobs such as deep research and scheduled automation, Google Workspace and email work, presentations, Canva design work, live dashboards, data cleanup and analysis, document analysis, finance and market monitoring, bug testing and code audit, and creative media (video, image, music, narration). When Agent Intuition is on, Alfrada OS can scan available playbooks, open the relevant one, and use it as part of the plan.
That is one reason the system improves with use. It is not just remembering your words. It is learning and reusing better operating patterns.
What Good Early Usage Looks Like
Strong users converge on the same five habits:
- Provide constraints, not magic — Give the system explicit parameters, context, and boundaries. The tighter the frame, the better the output.
- Name your deliverables — Always specify the exact output format: Markdown memo, structured CSV, 10-slide deck. Do not leave it vague.
- Ask for continuity explicitly — Tell Alfrada OS to search history and pull the relevant context forward when the job started earlier.
- Separate chat from artifacts — Keep chat for steering and revision. Let the Work Panel hold the finished assets.
- Codify your success — When a workflow works well, save it as a Playbook so you never redesign it from scratch.
Beyond those principles:
- ask Alfrada OS if it can help you achieve the objective before worrying about tools or setup
- attach context before asking for analysis
- let it use playbooks or prior workflows when the job is familiar
- refine the output in follow-up turns instead of restarting
Trust The Agent, Verify The Important Parts
Alfrada OS is designed to reduce a lot of the usual friction that comes with language models by combining planning, tools, memory, files, and working context in one system.
That does not mean normal LLM failure modes disappear completely. It means they are reduced and better contained. For high-stakes work, important facts, numbers, and decisions should still be checked when in doubt.
The right posture is simple: trust Alfrada OS to propose the path forward, but fact-check where the cost of being wrong matters.
Private By Design
Alfrada OS is designed for serious work, which means trust matters as much as capability.
Your chats run on a mix of EU and global model providers, with Zero Data Retention enforced wherever it is available — the company running the model deletes your prompt and its answer as soon as the response is produced. If your work has to stay in Europe, an EU Data Residency mode restricts every model choice to EU-hosted models. The full picture — including the automatic privacy fallback and the one deliberate exception — is in Privacy And Data Routing.
Two Things You Will Notice Early
The Model Ceiling On New Free Accounts
Brand-new Community accounts start with access to every model up to the Workhorse 3× band; higher-band models show as locked in the picker with an explanation. The ceiling lifts after a quick account review — normally within a day or two of signup — or immediately after any purchase, even a small credit pack. Paid plans are never capped. This is abuse protection for shared capacity, not a judgment about you — and the same ceiling applies to Auto, Swarm workers, and scheduled automations, so an expensive model can't be reached through a side door.
The Effort Selector
In the model selector on the message box there is a second row: Effort — Auto / Fast / Medium / Beast. It controls how thoroughly Alfrada OS double-checks its own work, separate from which model you choose. Fast skips self-review entirely and is the fastest. Medium does one self-review-and-revise pass. Beast reviews and revises until its own check passes, and also grades any generated charts and images. Leave it on Auto and Alfrada OS picks the best model and effort for each turn. More in Models And Modes.
Account Basics Worth Doing Once
- Turn on two-factor authentication. The post-login "Protect your account" prompt is Alfrada's own 2FA offer, and it is safe to accept. Enrollment works with any authenticator app at Settings → Account → Security; you get 10 one-time backup codes, a trusted device can skip the code for 30 days, and the prompt can be snoozed if now is a bad time. Full walkthrough in Account Security.
- Know how sharing works before you need it. Any conversation can be shared by private email invite or public link. Before anyone sees it, tool runs, system prompts, and the memories Alfrada OS brought into the conversation are stripped out — recipients see the clean narrative, and can claim their own copy to continue the work. Manage every share at Settings → Shares. Details in Sharing Conversations.
- Set the basics in Appearance. Settings → Appearance holds the comfort settings: a Light, Dark, or System theme; the layout — Book View opens the Work Panel automatically beside the chat, Page View keeps it tucked away until you open it; and how Alfrada OS addresses you (title and gender, used in voice messages).
Where To Go Next
- Quickstart if you want the fastest path to a first useful result
- Writing Better Prompts if you want better prompts and copyable examples
- How Alfrada OS Works if you want the deeper product model
- Playbook Library if you prefer learning from concrete use cases