Can You Choose Which AI Model Makedraft Uses?

Updated

Yes, on Makedraft's Pro and Business plans you can choose the AI model that fits the work, and you can switch models on a chat that has gotten stuck. The site does not list which specific models are available, so treat the exact lineup as something to confirm in the product before you commit. On lower-tier plans, the site's wording implies model choice is not offered.

How model choice works

Makedraft describes itself as an AI builder that produces real code in a real GitHub repo rather than a throwaway sandbox preview. Each chat opens a branch, and each commit on that branch deploys automatically as its own live service.

Model choice sits on top of that pipeline:

  • Plan-gated: The site states that Pro and Business plans "aren't locked to a single model." That phrasing is a plan feature, not a universal one.
  • Task-based selection: The stated purpose is to "choose the one that fits the work" — for example, a model better suited to a tricky UI component versus one better at backend logic.
  • Mid-chat switching: A testimonial on the site describes switching the model on a chat that was stuck instead of starting over. The site presents this as a supported behavior, not a workaround.

What stays the same when you switch

Because every chat is already a branch in your repo, changing models does not change where the output lives. The code, commits, and deploys remain in your GitHub repository and on your Render-hosted service. Switching models changes the assistant generating the next changes, not the destination of the work.

What the site does and does not tell you

Question What the site supports
Can I pick a model? Yes, on Pro and Business plans
Can I switch mid-chat? Yes, per the site's own account of a stuck chat
Which models are offered? Not listed on the page
Is model choice on every plan? No — the site ties it to Pro and Business
Does switching break my deploy? No indication of that; deploys are tied to commits, not models

The gap is the model list itself. If a specific model matters to your workflow, check the pricing page or the product before signing up rather than assuming a particular provider is included.

Why this matters in practice

Model choice is most useful when a task stalls. A common pattern with AI code builders is that a chat reaches a state where the assistant keeps producing changes that don't fix the problem. On a single-model tool, the usual escape is to start a new chat and lose context. On Makedraft, the site's account is that you can change the model on the existing chat and continue.

A concrete scenario: you're building an internal dashboard and the agent keeps misreading a data-fetching pattern. Rather than rebuilding the branch from scratch, you switch the model for that chat and re-prompt with the same context. Your repo history and live deploy stay intact.

Choosing between plans on this basis

  • If model flexibility is a requirement, the site points you to Pro or Business. Budget for one of those tiers rather than assuming the entry plan includes it.
  • If you mainly want the repo-first workflow — real branches, automatic deploys, your own domain, React/Next.js projects — model choice may be secondary, and the lower tier may be enough.
  • If a specific model is non-negotiable, verify availability before purchasing. The page confirms the capability but not the catalog.

For the surrounding workflow — how a chat becomes a deployed app, and how this differs from sandbox-based builders — see the related questions on Makedraft's chat-to-deploy pipeline and its repo-first architecture.