How to Use ReadyHermes: Hermes Agent Setup Guide

Start your free trial

The trial gives you 30 minutes of the Lite plan. No card is needed. The clock starts when your agent first runs, not when you sign up, so you will not lose time on the setup screen.

1. Open readyhermes.com and click “Start free trial”.
2. Sign up with an email and password, or choose “Continue with Google”. You can also go straight to the sign-up page.
3. A setup screen shows progress while your agent starts. This takes about a minute. Leave the tab open.
4. The dashboard opens on the Assistant tab. Your agent is running and ready to talk.

Each account gets its own dedicated Hermes Agent, with its own memory and skills. It is not a shared chatbot, and other users cannot see yours.

When the 30 minutes end, the agent pauses. You then have 24 hours to choose a plan and keep it, with its memory and files. If you do nothing in that window, it is deleted. If you plan to connect apps and test schedules, do it in one sitting.

Have your first chat

The Assistant tab is a chat box with a few suggested starter prompts. Click one, or type your own message. You can use it right in the browser.

The left menu has two groups. Under Workspace you will find Assistant, Channels, Prompt library and Conversation history, plus Usage and Notifications. Under Manage you will find Billing, Model key, Bot tools, Assistant settings, Security, Account, Privacy & data and Help & support.

Good first asks are specific and have a clear result. A few that work today:

• “Compare the three cheapest CRM tools for a five-person team, as a table.”
• “Every weekday at 8am send me the five biggest moves in my industry.”
• “Remind me on Friday at 4pm to chase the invoice.”

The agent has long-term memory, so tell it things worth keeping: your role, how you like answers formatted, the tools you use. It also has web search, a real web browser and its own Linux computer, and it can write skills for itself over time.

The Prompt library has ready-made starting points for planning, writing and decisions. Conversation history shows the conversation saved to your account, so you can pick it up on another device.

Connect a chat app

Browser chat is fine at your desk. To reach your agent from your phone, and to get scheduled results pushed to you, connect a chat app. Open Channels in the menu. Most connections take about a minute, because the assistant restarts while it picks up the change; the first WeChat connection takes 2 to 4 minutes.

Telegram

1. In Telegram, open a chat with BotFather.
2. Send /newbot and follow the prompts to name your bot.
3. Copy the token BotFather gives you.
4. In the dashboard, go to Channels, then Telegram, then Connect. Paste the token.
5. Message your new bot in Telegram. It replies with a short pairing code.
6. Paste that code into the dashboard and click Approve.

From then on, the bot replies to you there.

WhatsApp

1. Go to Channels, then WhatsApp, then Connect. A QR code appears.
2. On the phone that will host the bot, open WhatsApp, then Settings, then Linked devices, then Link a device.
3. Scan the QR code.
4. Keep that phone online.

This uses WhatsApp’s linked-device web session, not the official Business API. Upstream Hermes recommends a dedicated number for the bot, and we agree.

Discord

1. In the Discord Developer Portal, open your application, go to Bot, and click Reset Token. Copy the new token.
2. Paste the token under Channels, then Discord, then Connect.
3. Invite the bot to your server.
4. Message your bot. It replies with a short pairing code; paste it into the dashboard and click Approve.

WeChat

1. Go to Channels, then WeChat, then Connect, and scan the QR code from the phone that will host the bot.
2. Wait. The first connection installs a plugin, so allow 2 to 4 minutes.
3. If the dashboard then asks for a pairing code, message your bot in WeChat, paste the code it replies with and click Approve.
4. Keep that phone online.

Feishu / Lark

1. In the Lark developer console, open your app and go to Credentials & Basic Info.
2. Copy the App ID and the App secret.
3. Enter both under Channels, then Feishu / Lark, then Connect.
4. Message your bot in Lark. It replies with a short pairing code; paste it into the dashboard and click Approve.

Connect your other apps

Your agent can work inside the tools you already use. More than 1,000 apps are available, including Gmail, Google Calendar, Google Drive, Google Sheets, Notion, Slack, GitHub and Outlook.

1. In the chat, ask in plain words: “Connect my Gmail.”
2. The agent replies with a secure sign-in link.
3. Open the link and sign in to the app.
4. Go back to the chat and tell the agent to carry on.

Each connection belongs to your agent only. Once Gmail and Calendar are linked, you can ask things like “Summarise my unread email and add any meetings to my calendar.”

Schedule jobs, reminders and page watches

Scheduled jobs are one of the main reasons to use a hosted agent. It runs while your laptop is closed. You set them up by asking in chat.

• A briefing: “Every weekday at 8am send me the five biggest moves in my industry.”
• A reminder: “Remind me on Friday at 4pm to chase the invoice.”
• A page watch: “Check this product page every morning and tell me if it drops below $120.”

Results are delivered to a connected chat app. Browser chat has no push, so nothing will pop up in a closed tab. Connect Telegram, WhatsApp or another channel first, then set the schedule.

To check or change a job later, ask the agent to list your schedules, then tell it what to change or cancel.

Work with files

Your agent reads files with code and answers with the actual numbers.

Hand your agent a file

1. Upload it in Bot tools, then Files, or send it to your agent as a file in a connected chat app such as Telegram. A CSV, spreadsheet or PDF all work.
2. Ask your question: “Which invoices in this sheet are overdue, and by how much?”

The browser chat itself has no attach button, so use one of those two routes.

Files in Bot tools

Open Bot tools, then Files. You can browse your agent’s storage, upload, download, rename and delete. There is no in-browser editor, so to change a file, download it, edit it on your machine and upload it again, or ask the agent to make the change for you.

Storage is 10 GB on Lite and 40 GB on Pro.

Bot tools

Bot tools is where the practical controls live. These are the tabs.

Power

Start, stop or restart your agent. Restart is the first thing to try if it feels stuck.

Backup and Restore

Backup downloads your agent as a .zip. By default it holds only Settings and Personality and instructions; tick Conversations, Workspace files and memory, and Skills too if you want those in the file. Restore loads a backup back in. Managed settings, such as the AI model and your connected chat apps, are not restored, so reconnect those after. Bot tools works while your trial or plan is active, so download a backup before access ends.

SSH

1. In your own Tailscale account, create a one-time auth key.
2. Paste the key into the SSH tab. Your agent joins your private network.
3. Bot tools shows the exact command to run, in the form ssh root@<its name>.
4. Run it from a device on your Tailscale network.

No public port is opened on your agent.

API

Turn on an OpenAI-compatible endpoint and get a key. Use it from your own scripts and servers. It is meant for server-to-server calls, not for web apps running in a browser.

Desktop

Watch your agent’s desktop and browser live while it works.

My computer

1. Open the My computer tab. It shows a one-line command.
2. Run that command on your Mac, Windows or Linux computer.
3. Choose which folders the agent can see.

Access is read-only by default. Saving files and running commands are separate permissions, and both are off unless you turn them on. You can disconnect at any time.

Use your own model key

AI usage is included on both plans. The included model is DeepSeek V4.1 Flash, with fair-use limits. If you would rather use another model, paid plans let you bring your own key.

1. Open Model key in the Manage menu.
2. Pick a provider, such as OpenRouter, OpenAI, Anthropic, Google Gemini, DeepSeek, Groq, Qwen, Z.AI, Moonshot or MiniMax. Any OpenAI-compatible endpoint also works, as long as it has a public https address.
3. Paste your key. For another endpoint, also enter its Base URL.
4. Pick a model from the list (Load models fills it in), or type the exact model name your provider uses.
5. Click “Check and switch”. ReadyHermes first sends one tiny test message with your key, which your provider bills you for, then your assistant restarts on your key. That takes about a minute.

After that, your provider bills you directly for usage. You can switch back to the included AI at any time. This option is not available during the free trial.

Plans and billing

Billing is monthly and you can cancel any time. Prices exclude local tax where it applies. Pro is the same agent resized in place, so you can switch between plans whenever you like. Both plans include the same AI.

1. Open Billing in the Manage menu and choose Lite or Pro.
2. Pay through the checkout. Payments are handled by Waffo, the merchant of record.
3. To change plan, update your card, get invoices or cancel, click “Manage billing”. This opens the Waffo customer portal. Sign in with the email you paid with.

If you cancel, the agent keeps working until the end of the period you paid for, then stops. It is deleted 7 days later unless you subscribe again. See pricing for the full details.

Privacy, help and support

Your account and agent data are stored in the EU. To write a reply, your messages go to the AI model provider, which may be outside the EU. ReadyHermes does not train AI on your chats and does not sell your data. The details are in the privacy policy.

The Privacy & data page lets you download a copy of your data or delete your account yourself, with no email needed.

Help & support has guides and common questions. If you are still stuck, email support@readyhermes.com or use the contact page.

Curious how this compares with running Hermes Agent yourself? Read self-hosting vs managed hosting, or see what managed hosting covers.