Project Settings

Project Settings control how a specific project is configured, executed, and connected inside Origin. While account-level settings manage identity and authentication, this area defines how agents behave within a single repository.

These settings apply to all activity inside the project, including:

• AI Task Discovery
• Task execution and code edits
• Pull request creation
• Chat and Code mode sessions
• Sandbox-based runs

They do not directly modify your repository. Instead, they define the operational rules under which agents access code, execute changes, and use resources.

The left sidebar for a project includes:

• Workspace
• Project Details
• Settings
• Models
• Secrets

Each section addresses a different layer of project behavior.

Workspace

The Workspace page shows the live execution environment backing the project.

At the top, you can see:

• Current workspace status (e.g., Running)
• CPU usage
• RAM usage
• Disk usage
• Restart, Stop, and Refresh controls

This is where you monitor the environment that runs agent tasks and sandbox operations.

For example, if a large dependency installation is triggered during a task, you may notice a temporary increase in CPU or memory usage. If something becomes unresponsive, you can restart the workspace directly from this page.

The Startup Logs section provides real-time logs of environment initialization. This is useful when:

• Verifying that the correct repository has been cloned
• Confirming services started successfully
• Debugging workspace startup failures

Below that, Workspace Settings allow you to configure runtime behavior such as:

• Automatic updates
• Autostart schedules
• Autostop policies (e.g., shut down after 12 hours of inactivity)
• Provisioning parameters like CPU, memory, disk size, repository branch, or template source

For example, if your project runs heavy build processes, increasing memory or CPU allocation here ensures smoother execution. If the project is used intermittently, enabling autostop conserves resources.

Provisioning changes typically require a rebuild, since they alter the environment configuration itself.

Project Details

The Project Details section defines the identity and metadata of the project.

Here you can manage:

• Project name
• Description
• Assigned team
• Milestones
• Cycles
• Project metrics (API Key, Project ID, Total tasks)
• Repository association
• Creation date

This section is often reviewed when:

• Clarifying ownership between teams
• Setting up structured planning through milestones
• Organizing time-based cycles for sprint tracking

For example, if you are running a feature development sprint, you can define a cycle and associate tasks with it. Milestones help group work around larger deliverables, such as “v1 Launch” or “Refactor Phase 2.”

At the bottom, you can:

• Archive the project (hide it while preserving data)
• Move the project to Trash

Archiving is useful for completed initiatives that may still need to be referenced later. Moving to Trash removes it from active use but allows recovery from the Trash section if needed.

Settings (Project Preferences)

The Settings tab controls how agents behave inside this project.

You can configure options such as:

• Auto-PR – Automatically create pull requests for new features and fixes
• Memories – Allow agents to retain previous conversational context
• Wiki – Enable project-level documentation
• Semantic Search – Improve contextual retrieval across the codebase

For example:

• If you prefer a review-first workflow, you may enable Auto-PR so every agent-driven change is surfaced as a pull request rather than applied silently.
• If you frequently iterate on the same feature over multiple sessions, enabling Memories helps agents maintain context without restating everything.
• Enabling Semantic Search improves how agents locate relevant files in larger repositories.

These controls shape execution style and workflow patterns rather than core capabilities.

Models

The Models tab defines the default AI model for the project.

You can:

• View available models
• Search and compare models
• Select a default model (e.g., GPT-4o or others)

The selected model determines how agents reason, generate code, and analyze the repository by default.

For example:

• A lighter model may be sufficient for small bug fixes or documentation updates.
• A more advanced reasoning model may be preferred for architectural refactors or multi-step task planning.

The model choice directly affects response quality, reasoning depth, speed, and cost. It becomes the default across chats and tasks unless explicitly overridden.

Secrets

The Secrets section manages environment variables and API keys used during execution.

You can:

• Add individual secrets
• Import from a .env file
• Filter by environment
• View update timestamps

Secrets are marked as sensitive and cannot be displayed after saving. They are encrypted and only injected into the execution environment when required.

For example:

• If a task requires calling an external API, the API key is stored here.
• If the project integrates with a database or KV store during sandbox runs, connection credentials are added as secrets.

Secrets are never written to the repository, committed to Git, or exposed in diffs.

How Project Settings Are Typically Used

Teams usually visit Project Settings when:

• Setting up a new repository connection
• Adjusting workspace resource limits
• Enabling automated pull requests
• Selecting a default AI model
• Configuring secrets for integration tests
• Organizing work into milestones and cycles
• Archiving completed projects

In short, this area defines the operational environment of the project. It ensures that agents run with the correct repository access, compute resources, automation rules, and security boundaries, while leaving the source code itself unchanged until tasks explicitly generate updates.