Generating Plans
Turn a completed brief into a structured task breakdown with one click.
Overview
Plan generation reads the content of your brief and produces a two-level task hierarchy: a set of parent tasks covering the major areas of work, each automatically elaborated into subtasks with detailed descriptions and acceptance criteria. The entire process runs in the background, and tasks appear in the plan view as they are created.
How It Works
-
Open your brief — Navigate to any brief in your workspace. The Plan tab shows an empty state until a plan has been generated.
-
Click "Generate Plan" — Hamster Studio analyzes your brief content and searches for relevant context from your connected blueprints, methods, and codebase. A progress indicator shows the current generation phase.
-
Parent tasks are created — The AI produces a set of top-level tasks that represent the main areas of work. These appear in the plan view as soon as they are ready, before elaboration is complete.
-
Subtasks are elaborated in parallel — Each parent task is expanded into detailed subtasks simultaneously. Each subtask includes a description, scope boundaries, and acceptance criteria. The plan fills in progressively as elaboration threads complete.
-
Plan is finalized — Once all tasks have been elaborated, the progress indicator clears and the full plan is available for review and delivery.
Key Capabilities
-
Live task streaming: Tasks appear as they are generated rather than all at once, so you can begin reviewing early parent tasks while the rest of the plan is still being built.
-
Context-aware generation: Before creating tasks, the AI searches your blueprints, methods, and codebase. This means the plan reflects your team's actual patterns and existing work, not just a generic breakdown.
-
Scope enforcement: Each parent task is elaborated with awareness of its sibling tasks. The AI is instructed not to create subtasks that overlap with work covered by other parent tasks, reducing duplication.
-
Optional focus areas: When triggering generation from the brief's AI conversation, you can specify a target number of tasks or particular focus areas to shape the output.
-
Iterative chat refinement: After a plan is generated, the plan view has an AI chat panel scoped specifically to your plan. You can ask it to adjust tasks without regenerating from scratch — for example, to split a parent task, add a missing subtask, rewrite acceptance criteria, or change the scope of a specific task. The AI reads the current plan state and applies changes directly.
-
Regeneration: If the generated plan does not reflect what you need, you can regenerate it from the Plan tab. Regenerating clears all existing tasks for the brief and produces a fresh plan from the current brief content. This action cannot be undone.
Generation Phases
The progress header shows the current phase as plan generation runs:
| Phase | What is happening |
|---|---|
| Analyzing | The AI is reading and parsing the brief content |
| Creating core tasks | Parent tasks are being created one by one |
| Extracting task list | The completed parent tasks are collected for elaboration |
| Creating subtasks | Each parent task is being elaborated in parallel |
| Finalizing | Elaboration threads are completing and the plan is being confirmed |
Tips
-
Write a thorough brief before generating a plan. The richer the context in your brief — requirements, constraints, background research — the more accurate and detailed the generated tasks will be. Use the Guided Research flow first if you need to gather information before generating.
-
If you are not happy with the initial plan, use the Regenerate button after editing your brief to add more detail or clarification rather than manually reworking every task.
-
Generation can take a few minutes for complex briefs. You do not need to stay on the page — the plan is persisted and will be ready when you return.