Skip to main content

​
Overview

Sometimes you want to explore different approaches to the same problem, or create several related tasks at once. Multi-try runs the same prompt multiple times in parallel, while batches let you create a set of distinct tasks or todos in one operation.

​
Multi-try

Multi-try launches the same prompt as multiple independent tasks (up to 4), each running in its own container. Because AI agents are non-deterministic, each attempt may take a different approach — letting you compare and pick the best result.

​
Creating a multi-try

  1. Write your prompt in the composer
  2. Click the multi-try button in the toolbar to cycle through attempt counts (1 → 2 → 3 → 4 → back to 1)
  3. Launch the task
Whim creates the specified number of tasks, all sharing the same prompt, provider, model, and source branch. Each task gets its own isolated branch and container.
Multi-try is especially useful for problems with multiple valid solutions — like UI implementations, algorithm choices, or architectural decisions. Compare the approaches and keep the one you like best.

​
Comparing results

Multi-try tasks appear as a group in your workspace. You can:
  • Open tasks side by side using split or grid layouts to compare output
  • Review the code changes on each task’s branch
  • Complete the attempts you don’t want and keep the winner
  • Create a pull request from the best result

​
Batch creation

Batches let you create multiple distinct tasks or todos in a single operation. Each item in a batch gets its own prompt and settings, but they all share a common relationship to their parent.

​
When to use batches

  • Decomposing work: Break a large initiative into specific child tasks
  • Parallel workstreams: Launch several independent features at once
  • Backlog seeding: Create a set of todos for a sprint or project

​
Batch settings

When creating a batch, all items share:
  • Kind — whether to create running tasks or backlog todos
  • Relationship — whether items are children of a parent task or top-level items
  • Parent — an optional explicit parent task for all items
Each item in the batch can have its own:
  • Title and prompt
  • Priority and effort
  • Tags
  • Assignee
Batches are limited to 10 items per operation. Each item follows the same creation flow as an individual task or todo.

​
Use cases

Multi-try

  • A/B testing different implementation approaches
  • Exploring UI layout alternatives
  • Finding the best algorithm for a performance problem
  • Getting multiple perspectives on a complex refactor

Batch creation

  • Sprint planning — create todos for each story
  • Feature decomposition — break an epic into tasks
  • Parallel development — launch independent features simultaneously
  • Onboarding — seed a new workspace with starter tasks