Apollia Desktop, help center
Step-by-step guides for setting up, automating and keeping control of your AI day to day.
This corpus covers Apollia Desktop, the desktop application, from the operator's side: install it, connect a model, hand tasks to agents, automate work, keep a hand on sensitive actions, and follow what happens.
Where to start
- Install Apollia Desktop, from the download to a first agent at work, on macOS, Windows or Linux.
- Getting acquainted, the few steps Apollia opens with the first time. The application then opens on the Chat.
- Set up your profile, the initial flow.
- Connect a remote model, add the AI server of your organisation: any OpenAI-compatible server.
The main blocks
- Get going : connect a model, install an agent, hand it a task, what became of projects.
- Talk : free conversation and voice dictation.
- Automate : schedule a trigger, follow the firing history.
- Keep control : approve or refuse an action, review what each agent may do, inspect a tool, choose an autonomy tier, measure an agent with eval.
- Connect : link your accounts, folders and tools on the Connections screen, or connect an MCP server.
- Follow : activity timeline, LLM costs, activity log.
- Maintain : update Apollia, find your version and your data.
- When something jams : diagnosing the common cases, a silent AI, a stuck agent, a refused action, dictation that produces nothing.
How to read this help center
Every page follows the same shape:
- Prerequisites, what you need before you start.
- Steps, numbered interface actions with screenshots.
- Verification, how to confirm it worked.
- When it does not work, the frequent error cases.
When a page mentions a technical concept, the link goes either to the Explanation (how it works) or to the Reference (the exhaustive spec). One jump, to the right place.
Going further with your agents
- Choose an autonomy tier, adjust how far an agent goes on its own before asking for your agreement.
- Measure an agent's performance with eval, build a reproducible test suite and read the results report.
A developer rather than an operator?
This help center is about using the application. If you want to build packages, agents with Python tools and workflows, or understand the internal architecture:
- the tutorials, learning to build agents step by step.
- the technical reference, CLI, HTTP API, SDK contract, configuration.