Make your first hour in Auxot count

Run one ordered pass through **System Health**, **Chat**, provider reality for admins, a tiny **Context Files** win, and a five-second **Audit Logs** peek so your first hour has checkpoints instead of aimless clicking.

Plus: three Admin Agent prompts that reorder the same five blocks for owners, members, and builders when time or permissions do not match the default script.

Audience Everyone · Admins
Time ~6 min
Prerequisites You are past **onboarding-blocked** for your org. Helpful: you already clicked the shell once ([Find the right screen for your next question](/tutorials/find-the-right-screen-for-your-next-question)). Owners mid-wizard: finish [Finish first-run onboarding](/tutorials/finish-first-run-onboarding) until **Chat** is reachable before you treat this as your first *productive* hour.
You'll end up with Five concrete checkpoints finished or honestly skipped with a note, a sense of which screen owns models versus memory versus history, and three Admin Agent prompts you can reuse when your calendar shrinks the hour to twenty minutes.

When a tutorial shows italic text in quotation marks, it usually mirrors a label or helper string inside Auxot. Product copy changes between releases — if something reads differently in your workspace, trust what you see on screen.

Callouts with a Worth knowing gold accent are meant as must-read context before you move on. Blockquotes that open with Tip are lighter, optional depth.

Why this matters

The left menu is not a todo list. Without an order, people burn an hour reopening Chat, rereading empty Context Files, and wondering why Slack still cannot tell which Auxot user is at the keyboard. The sequence below mirrors how operators actually trust a stack: health, conversation, configuration, durable memory, then evidence.

You are not trying to finish Auxot in sixty minutes. You are trying to leave with one working thread, one honest read on providers (or a clear ping to whoever owns them), one fact your agents can reuse, and one mental bookmark for where jobs land in Audit Logs.

Today, you run the five steps in order and stop when the hour ends. The next time you open Auxot, you run the same five steps in about ten minutes because you already know what each one is for.

Members may skip provider steps. Admins may skip context if the fire drill is routing. Everyone still benefits from System Health plus Chat first.


Quick start

  1. Skim System Health: click System Health in the left menu. Spend about sixty seconds on cards and colors only (Take Auxot’s pulse in 10 seconds).
  2. Say hello in Chat: open Chat, pick the Admin Agent, send hello, and wait for a streamed reply (Say hello to the Admin Agent).
  3. Check provider reality if you admin the org: open Settings → Providers long enough to see whether a cloud or worker path already exists (Connect a cloud AI model). If you are not an admin, skip this step and note who owns providers instead of guessing.
  4. Add one small context win: open Context Files and create the smallest file you will actually use: a title plus one paragraph you already trust from email, docs, or policy (Add your first context file).
  5. Peek Audit Logs once: open Audit Logs, scroll for five seconds, close the tab. You are learning where job history lives, not investigating an incident (View your audit logs).

Done? You can name, out loud, what you verified in each step and what you deferred with a written next action.


The agent can do that?

You ran the five steps once (or you hit a permission wall and took notes). These three prompts ask the Admin Agent to re-sequence the same blocks for your real constraints. Paste them in Chat → Admin Agent.

1. Reorder the hour for your role

I am [org owner / member / builder with API work]. I have about one hour. Reorder these blocks for maximum leverage: System Health, Chat with Admin Agent, Settings → Providers, Context Files, Audit Logs. Mark each block as do-now, do-later, or skip-with-reason. Cap the answer at eight bullets.

Why it’s non-obvious: Owners think everyone should touch providers. Members think Chat fixes Slack. A forced ordering surfaces what your role should actually own in hour one.

2. Shrink the hour to twenty minutes without shame

I only have twenty minutes left in my first Auxot session today. Given what I already completed from System Health, Chat, Providers, Context Files, and Audit Logs, list the smallest subset that still earns the session, what I should explicitly defer, and one line I should write in my notes so future-me does not redo work.

Why it’s non-obvious: Partial progress beats heroic scope. The leverage is naming what you refuse to start so you do not leave three half-open tabs that feel like failure.

3. Draft a tight ping when you are blocked on someone else

Draft a three-sentence message I can send my Auxot org admin. Facts: I finished System Health and Chat, but I cannot open Settings → Providers / billing / invites. Neutral tone, no blame, asks for one specific action or ETA. Mention [Connect a cloud AI model](/tutorials/connect-a-cloud-ai-model) or [Finish first-run onboarding](/tutorials/finish-first-run-onboarding) only if relevant.

Why it’s non-obvious: Stuck teammates either go silent or write novels. A three-sentence ping moves the owner without sounding like a ticket queue.


Go deeper

When Linked Accounts belongs in hour one instead of Context Files

If your team already talks in Slack or Discord this week, swap the depth slot: open Settings → Linked Accounts and finish the link before you invest in Context Files (Link your chat apps to your Auxot account). Identity gaps make every later step feel broken.

When API Keys belongs in hour one instead of Context Files

If you already know you are wiring a script, GitHub Action, or intake today, spend your depth slot on Settings → API Keys instead (Generate your first API key, then Pick personal or team API keys for real work). Do not create keys you will not store safely before you leave.

If Chat never streams a reply

Treat it as a provider or onboarding gate, not a mystery bug. Return to Unstick your first week in Auxot, then reopen System Health before you file support.


Walkthrough

Step 1: System Health as the clock start

Open System Health before Chat. If a provider or Redis is red, the rest of the hour is not time well spent. You are not fixing everything now; you are deciding whether to keep going or to flag the red card and stop.

Step 2: Chat as the human proof point

Open Chat, pick the Admin Agent, send hello. Streaming text is the proof that your account can reach a model route end to end. If the picker is wrong or empty, fix that before you build agents.

Step 3: Providers (admins) or a deliberate skip (everyone else)

Admins: Settings → Providers. Confirm at least one sane path (cloud or worker) matches what your org expects. Members: send the short ping you drafted in The agent can do that? instead of clicking in circles.

Step 4: Context Files as the smallest durable win

Open Context Files. Create one short file with information you already stand behind: pricing bullets, support tone rules, or a product one-pager. Add depth later; hour one is about proving the Context Files page works for you.

Step 5: Audit Logs as the closing peek

Open Audit Logs. Scroll once. You are confirming every job and event Auxot runs is recorded here, with timestamps and the data needed to investigate later, so week-two-you knows where to look instead of treating Auxot as opaque.


What’s next

Reference