AI models and routing in Roseram
Roseram can route tasks across configured frontier models or let you pick a model when you need control.
Automatic Routing
Automatic Routing chooses a model profile for the job — for example a stronger coding model for refactors vs a faster chat model for framing.
Availability depends on configuration and capacity. Plus and Pro differ in processing speed, not in “locked out of models.”
Manual picks
Pin a model when you want reproducibility (evals, demos, support repros). Name the model in your thread notes so teammates know what won.
Practical defaults
- Plan / brainstorm — faster frontier chat profile
- Build / tool loops — stronger coding profile; Pro lightspeed helps
- Review — careful model + human gate on diffs
Workflow ·
More Roseram workspace guides
Getting started with Roseram
Open a project, write a brief, and review your first AI changes.
Workspace troubleshooting
Find help for previews, connections, and failed requests.
Plan, build, and review code
Choose the workspace mode that fits your next task.