Start tasks with AIRI loci Agent
Working with AgentReview draft
On this page
AIRI loci Agent is your AI partner inside a project. Organize references, research, discuss designs and generate images or video in the same conversation, without switching image/video modes.
Define this round
Attach relevant inputs, then state the goal, fixed conditions and expected output. Keep default models and parameters so Agent can select suitable tools.
| Goal | How to start |
|---|---|
| Understand a brief | Upload it; ask for requirements, constraints and questions |
| Research and compare | Provide precedents, comparison criteria and a request for sources |
| Generate or edit images | Reference the base design and inspiration; specify viewpoint, atmosphere and fixed elements |
| Generate video | Supply inputs; describe the scene, camera movement and stable elements |
| Follow a method | Select a Skill and provide its required inputs and goal |
If the direction is uncertain
Discuss and compare before deciding to generate: “Analyze the brief's constraints and suggest two directions; do not generate images yet.”
Explicitly reference the relevant file or image when continuing work, rather than “the last one” or “that scheme.” Follow any prompts about attachment types, counts or combinations.