​
1. Agent Options (UI Phase)

​
Overview

Codiris agents adapt to the complexity of your design task. You can select the right “thinking level” for your project phase.

​
Agents

  • Fast
    • Best for small tweaks or rapid refinements.
    • Lightweight computation for speed.
    • Example: “Change the button color to blue.”
  • Medium Thinking
    • Balanced between speed and depth.
    • Ideal for full websites or medium-complexity changes.
    • Example: “Generate a landing page with signup, pricing, and blog.”
  • High Thinking (available for paid users)
    • Deep reasoning for complex foundations.
    • Builds strong design systems and multi-layered apps.
    • Example: “Create the first version of a SaaS dashboard with role-based access.”

​
Recommendation

Start with High Thinking for a solid base → Switch to Fast for refinements → Use Medium for full websites.

​
2. Multi-Agent System in Codiris

Codiris doesn’t rely on just one AI model it uses a multi-agent ecosystem, where each agent specializes in a specific role (designer, researcher, coder, tester, PM, etc.). These agents collaborate, hand off tasks, and even debate solutions to reach the best outcome. lab9.png
  • Specialization → Each agent is tuned for a domain (e.g., Research vs. UI vs. Coding).
  • Parallelization → Multiple agents can work at the same time to accelerate delivery.
  • Checks & Balances → Agents critique each other’s output, reducing errors.
  • Human-like workflows → Mimics how real product teams (PM, designers, engineers) collaborate.

​
🧩 Types of Agents

  • Research Agent → Collects market data, competitor insights, trends.
  • Design Agent → Creates wireframes, branding kits, UI mockups.
  • Coding Agent → Writes, tests, and optimizes production-ready code.
  • Testing Agent → Generates unit/integration tests, runs QA scripts.
  • Deployment Agent → Handles CI/CD, pushes builds live.
  • PM Agent → Organizes user stories, prioritizes backlog, manages versions.

​
🔄 Collaboration Modes

  1. Sequential Mode
    • Agents pass work to each other in order.
    • Example: Research → Stories → UI → Coding → Testing → Deployment.
    • Best for structured product development.
  2. Parallel Mode
    • Multiple agents work simultaneously on different aspects.
    • Example: Coding Agent writes the code while Design Agent refines the UI.
    • Best for speed and iteration. image.png

​
🚀 Benefits of Multi-Agent Workflow

  • Faster delivery – Work streams happen in parallel.
  • Higher quality – Agents cross-check each other.
  • More creativity – Brainstorming agents bring diverse approaches.
  • Less manual overhead – Agents organize themselves like a self-managed team.

​
3. Live Calls

Codiris includes built-in video/audio calls with collaborative editing and screen sharing — making it easy to brainstorm, review, and build together without leaving your workspace.
  • Video & Audio
    • High-quality calls integrated directly into Codiris.
  • Collaborative Editing
    • Edit prompts, designs, and tasks live during the call.
    • Everyone sees changes instantly.
  • Screen Sharing
    • Share your screen to walk teammates through designs, code, or dashboards.
    • Perfect for design reviews, live debugging, or demoing features.
    • Multiple participants can share in the same call.
  • Agent Participation
    • Bring Codiris agents into the call.
    • Agents can present research, generate UI updates, or suggest code fixes in real time.
  • Context-Aware Discussions
    • Calls are tied to your active project workspace.
    • Agents and teammates always see the same context.