Campaigns

Campaigns let you group related conversations into a single organizational unit within Ivy. Use them when several chats share a common objective -- a research project, a drug target investigation, or a screening run -- and you want them organized together instead of scattered across your chat history. Campaigns sit one level above individual chats, giving your discovery work a lightweight project structure.

Campaigns

Campaigns let you group related conversations into a single organizational unit. A campaign might represent a research project, a drug target investigation, a screening run, or any logical collection of chats that share a common objective.

Campaign hierarchy

Campaigns live inside projects, and chats live inside campaigns:

Project
  └── Campaign
        ├── Chat 1  (e.g., "BRCA1 target validation")
        ├── Chat 2  (e.g., "Lead compound docking")
        ├── Chat 3  (e.g., "ADMET profiling of top hits")
        └── Chat 4  (e.g., "Synthesis route planning")

Every campaign belongs to exactly one project.

Each campaign contains one or more chats. A chat belongs to exactly one campaign (or no campaign if unassigned). Canvas cards, tool results, and exported files within a chat are associated with its parent campaign.

Creating a campaign

1. Open the project in the file explorer.
2. Click New Campaign in the project's quick actions.
3. Enter a name and optional description for the campaign.
4. The campaign is created immediately and appears under the project.

Adding chats to a campaign

Associate a chat with a campaign when you create it: start the new chat from inside the campaign in the file explorer, or pick the campaign in the new-chat dialog.

Campaign-level views

Chat list

Each campaign shows all its chats in chronological order. Click any chat to open it in the workspace. Pinned chats within a campaign appear at the top.

Job history

The Next Steps (Campaign) action in the action rail lets you view all jobs submitted from chats within a specific campaign. This gives a consolidated view of computational work across related conversations.

Canvas aggregation

Canvas cards generated across all chats in a campaign can be browsed together through the campaign content panel. This is useful for comparing results from different analyses -- for example, viewing docking scores from multiple chats side by side.

Renaming and editing campaigns

Open the campaign in the file explorer and update its name or description. Only the campaign owner can change it.

Deleting a campaign

Only the campaign owner can delete a campaign.

Campaign analytics

Campaign analytics are available on the Analyze pages. Navigate to /analyze/campaign/[campaignId] to see:

• Chat count -- Total conversations in the campaign.
• Job summary -- Breakdown of jobs by type and status (completed, failed, running).
• Tool usage -- Which computational tools were used most frequently.
• Timeline -- Chronological view of activity across the campaign.

Best practices

One campaign per research question

Create a separate campaign for each distinct research objective. For example, keep "Target Identification" and "Lead Optimization" as separate campaigns even if they share a drug program.

Use chats for individual experiments

Within a campaign, create a new chat for each experiment or analysis step. This keeps context windows fresh and makes it easy to revisit specific results.

Archive completed campaigns

When a campaign's research objective is met, rename it with a prefix like "[DONE]" to distinguish it from active work. Campaigns cannot currently be archived, so naming conventions serve this purpose.