Give it a purpose. Add useful skills. Bring your next question. Lomus hosts your Hermes agent so you can focus on what you want to do.
Studio is your agent’s home inside Lomus. Start conversations, attach supported files, and work with responses in one place. Studio conversations belong to their creator; they are not automatically shared with everyone in your workspace.
Talk with your agent on Telegram, Slack, or Discord, alongside Studio. The self-serve creation wizard currently exposes Telegram; contact support for help with Slack or Discord setup if those options are not shown. Each agent has at most one external channel. Conversations are separate; do not assume a shared transcript between chat apps and Studio.
Browse curated skills for research, writing, and other workflows. Enable them for your agent or upload a custom skill for safety review. Some require your own credentials or additional configuration; the catalog lists requirements. Review is a safeguard, not a guarantee of every skill’s output.
Select a supported OpenAI or Anthropic model in the app. Add instructions to shape the agent’s tone and approach. Use a supported provider key, or choose Lomus-managed credits where your plan allows it. The app shows the current model options.
See per-agent usage, model costs, activity, and available budget controls. Pause, resume, or restart agents as needed. Review conversations, generated files, and activity in the relevant app views while keeping an eye on spend.
Invite people by email to a shared workspace. Owners have full access including billing; admins manage the workspace and members; members can run, create, and delete agents; operators run existing agents. Work together with shared agents and skills, while Studio conversations remain personal to their creator.
Store API keys, passwords, and tokens in the secrets vault. Values are encrypted at rest and not returned by the management interface after saving. Shared secrets are available to authorised workspace use; private secrets are resolved for their owner. An agent can request a secret through an authorised runtime path, without needing you to paste it into a conversation.
Eligible plans include voice capabilities and image generation. These have their own configuration and usage requirements. They are capabilities within supported experiences, not standalone phone, SMS, or voice-call channels.
Start with one agent. Give it a purpose. Take it from there.
Create your agent Explore with a Trial workspace. Model usage costs are separate.