Purpose
ReadyHermes is a managed AI agent service. It deploys the open-source Hermes Agent developed by Nous Research on cloud servers, so users do not need to rent their own server, install Docker, or configure model keys; after registering an account, they can obtain a running dedicated Hermes Agent. According to official information, the agent is ready in about one minute, and each account has an independently isolated agent instance with its own memory and skills.
Main Capabilities
The agent has the following core capabilities:
- Long-term memory: Retains context across conversations and gradually accumulates information about the user's preferences and ongoing work.
- Self-learning skills: Can write reusable skills on its own based on experience.
- Scheduled tasks: Executes jobs according to a user-set schedule, for example sending an industry news briefing at 8 a.m. every workday.
- Real browser: Searches the web and opens pages in a real browser, returning results with sources.
- Built-in Linux environment: The agent has its own cloud computer and desktop, and users can at any time view its running screen and saved content in a browser tab, and browse its files.
The specific task types listed officially include:
- Morning briefing: Choose a topic, read the news, and send a short summary with sources to a chat app at a set time.
- Research and comparison: Search the web, open pages with a real browser, and return comparison results with sources (for example, comparing the three cheapest CRM tools in table form).
- Inbox and calendar: Connect Gmail or Google Calendar via a secure login link, for summaries, drafting replies, and adding events.
- Files and spreadsheets: Upload PDFs, CSVs, or spreadsheets, use code to read the files, and give answers with actual numbers.
- Web page change monitoring: Check prices, inventory, job postings, or your own website on a schedule, and send a message to a chat app when changes occur.
- Reminders and follow-ups: Explain once and the agent will send a reminder at the specified time through the connected chat app.
How to Use
- You can chat directly with the agent in the browser, or send messages from chat apps such as Telegram, WhatsApp, Discord, WeChat, and Feishu (Feishu/Lark).
- App connections: After making a request in chat, the agent sends a secure login link; each connection belongs only to that user's agent. The officially listed connectable apps include Gmail (read, summarize, draft emails), Google Calendar (view schedules, add events), Google Drive (find and read documents), Google Sheets (read and update spreadsheets), Notion (search, write, update pages), Slack (read and post), GitHub (follow up on issues and pull requests), Outlook (read and send emails), and more than 1,000 other apps.
- Model access: Each plan includes AI usage based on DeepSeek V4.1 Flash, no API key required, subject to fair use limits. Users can also paste their own API key (a test message will be sent first to verify before switching, and they can switch back to the built-in AI at any time), or connect to any OpenAI-compatible endpoint (requires a base URL, key, and model name, and requires a public https address).
- Connect a local computer: You can connect a Mac, Windows, or Linux computer and choose which folders are allowed (read-only or read-write optional); saving files and running commands are separate permissions, disabled by default; you can disconnect at any time, and after disconnecting the agent immediately loses access to these folders.
- Bot tools (located in the dashboard):
- Root SSH: Open a terminal to the agent machine from a local computer via Tailscale; SSH does not expose public ports.
- API: Enable an OpenAI-compatible endpoint for your own scripts and applications to communicate with the agent.
- Backup and restore: Save the agent as a .zip file at any time, and later restore it from the Restore tab.
- File management: Browse, upload, download, and rename agent files in the browser.
Use cases
Suitable for users who want to use Hermes Agent but do not want to deploy and operate it themselves, as well as individuals or small teams that need to complete information retrieval, schedule and email handling, file and spreadsheet analysis, web monitoring, and scheduled reminders via chat.
Limitations and notes
- Pricing: All accounts start with a 30-minute Lite free trial, no credit card required; the trial time starts when the agent first runs, and after it ends the agent is paused, and you need to choose a plan within 24 hours to keep memory and files, otherwise they will be deleted. Lite is $5/month (1 vCPU, 2 GB memory, 10 GB storage), Pro is $10/month (2 vCPU, 4 GB memory, 40 GB storage), prices exclude applicable local taxes, payment is processed by merchant of record Waffo, and you can cancel at any time.
- Data and privacy: Account and agent data are stored in the EU, and the agent runs in an isolated environment; to generate replies, messages are sent to AI model providers (which may be outside the EU); ReadyHermes does not use conversations to train AI models and does not sell data.
- Cancellation and deletion: After cancellation, the agent continues to work until the end of the paid period, then stops and is deleted after 7 days unless you resubscribe.
- Backup restore scope: Settings managed by the service provider (such as the AI model and connected chat apps) are not included in the restore scope.
- Trial limitations: During the trial, the Lite configuration is used; plans can be switched at any time from the billing portal, and the agent retains its memory, skills, and connected apps.
- Information on supported countries/regions, specific language support, platform client versions, and the specific details of fair use limits is not yet available.
