opayaQuestions,
answered.
Everything people ask before moving their agents into OPAYA. If something is missing, the demo is the fastest way to get an answer.
Book a demoWhat exactly is OPAYA?
A desktop workspace for AI agents. It installs, connects, updates, clones and backs up agents on your computer, on servers over SSH and in Docker, and brings their chats, terminals, projects and secrets into one window.
Which agents work inside it?
Hermes, Claude Code, Codex, OpenClaw, OpenCode and Goose, any OpenAI-compatible API such as OpenAI, Anthropic, Gemini, Groq, OpenRouter, Mistral, xAI, DeepSeek, Ollama or LM Studio, and custom CLIs. Interfaces and session features differ per agent.
Do I need an API key to start?
No. Start free installs Ollama and a free open model on your machine, then connects the built-in Opaya Agent. Free tiers from Ollama Cloud, OpenRouter, Groq, Cerebras, Gemini and Mistral work as well, and your existing agents keep their own accounts.
Which operating systems are supported?
OPAYA runs on Windows 10/11 (x64) and on Macs with Apple Silicon or Intel. The Mac preview is not notarized yet, so open it the first time with right-click then Open. After that it updates itself from inside the app.
Can OPAYA manage agents on servers and in Docker?
Yes. Add a VPS over SSH, and New VPS creates the key and checks the host for you. Install agents there as a regular install or a Docker container, clone an existing agent to it, open terminals, browse files, check versions and update everything. OPAYA never needs a server of its own.
How are API keys and secrets stored?
In an encrypted vault on your machine. Agents and saved chats see a reference such as [secret S1 · OPENAI_API_KEY · sk-p…9f3a], never the value. Clone and transfer copy credentials between agents without displaying them.
Does my code leave my machine?
Runtimes stay where you run them. OPAYA connects locally or over SSH; it does not route your work through its own cloud. Each model provider's own data handling still applies, and free local models keep prompts on the machine.
What happens when I update an agent?
OPAYA checks agents and their tools every hour on every machine and updates each one the way it was installed. OPAYA itself checks its releases, verifies checksums and installs in place; if an update fails it says so and offers the installer.
Can I run an agent on a schedule?
Yes. Schedules let an agent run without you driving the chat, for example a nightly check or a recurring report, on the machine where that agent lives.
What does cloning an agent actually copy?
Skills, MCP server configuration, secrets and memory, so the same agent behaves the same on a second machine. You choose the target host and OPAYA handles the transfer.
Can I use several agents on the same project?
Yes. Sessions are scoped per agent and per project, so a coding agent, a reviewer and a Hermes agent can work in the same repository without mixing history or drafts.
Does OPAYA replace Claude Code or Codex?
No. It runs them. You keep the agents and accounts you already have, and OPAYA adds the workspace around them: projects, git, terminals, machines, MCP servers, secrets and updates.
Do the agents see the same settings on every machine?
Project instructions live with the repository and tool configuration lives in OPAYA, so an agent started on a remote host follows the same rules as the one on your laptop. Model choice and reasoning effort are set per chat.
Can I reconnect to a session I closed yesterday?
OPAYA keeps agent-scoped history and drafts. Supported local sessions can run outside the interface, and remote terminal workflows can reattach to persistent shells. Exact behaviour depends on the runtime you use.
How do I try it without a sales call?
Install OPAYA and click Start free: no account and nothing to type. When you want a guided tour, the demo is 30 minutes with the people who build it.