简体中文
EnglishWebsite review
This review may include AI-generated content and is not individually verified. Domain, certificate and other facts come from the observed data on this page.
What is Makedraft?
Makedraft is an AI app builder that works directly in a GitHub repository. Instead of generating a throwaway preview you later have to export, each chat opens a real branch in your own repo, and every commit the agent pushes redeploys that branch as a live service automatically. It is aimed at people shipping real projects—indie developers, freelancers and early-stage founders—and works with React and Next.js projects, with custom domains and a choice of AI model on paid plans.
What that means in practice
The distinguishing feature is the absence of a migration step. A freelance designer, for example, can approve commits and have a client demo live a minute later without running a separate deploy, while an indie developer keeps full ownership of the code because it was in their repo from the first message. The trade-off is that Makedraft assumes you are comfortable with a Git-based workflow and a React/Next.js stack; someone wanting a drag-and-drop site builder with no repository involved is a poor fit.
| Consideration | Makedraft's approach | Who it suits |
|---|---|---|
| Where code lives | Your own GitHub repo, one branch per chat | Developers who want to keep or hand off the code |
| Deployment | Automatic on each commit | People who demo or ship frequently |
| Model choice | Selectable on Pro and Business plans | Teams wanting to switch models mid-project |
| Stack | React and Next.js | Standard front-end web apps, not arbitrary software |
The page also lists GitHub, Render, Supabase and Stripe as the underlying tools, and offers templates for dashboards, landing pages, components and small apps—useful if you want to see the workflow before committing a real project.
Next step: browse the templates and start one chat on a small internal tool, then confirm the branch and deploy behaviour match how your team reviews code. Pricing details are on Makedraft's pricing page.
How does Makedraft connect to a GitHub repo and deploy my app automatically?
Makedraft opens a real GitHub branch from your first chat message, then automatically redeploys each branch as its own live service whenever the agent pushes a commit. There is no separate sandbox or export step: the code you discuss is the code in your repo, and the branch is already deployed.
How the loop works, based on the page
- Start a chat — Makedraft connects that conversation to a real GitHub repository and opens a branch rather than a throwaway preview.
- Build in the branch — changes land as commits on that branch, so you can review them in GitHub like any other work.
- Commit triggers deploy — when the agent pushes, the branch's live service redeploys automatically. Each branch has its own deployment, which suits previewing a change before merging.
- Point a domain — you can attach a custom domain instead of using a shared subdomain, according to the page.
The page names GitHub, Render, Supabase and Stripe as the underlying tools, and says it works with React and Next.js projects. Pro and Business plans let you choose the AI model per chat, which matters if one model stalls on a task.
What this means in practice
For a freelancer preparing a client demo, the practical gain is that an approved commit is live shortly after, with no deploy step to remember. For an indie developer, the repo is the source of truth from the start, so there is no migration later. The trade-off is coupling: your workflow inherits GitHub branch habits and the hosting behavior of the connected service, so teams with strict release gates or a different CI pipeline should check how branch deploys fit their process.
A sensible first test is a small internal tool: connect the repo, make one change, confirm the branch deploys, then decide whether to merge or keep iterating on the branch.
Can I choose which AI model Makedraft uses for each chat?
Yes—on Makedraft's Pro and Business plans you can pick the AI model for a chat rather than being locked to a single one. The site frames this as choosing "the one that fits the work," and one quoted user (Sofia R., an early-stage founder) says switching models on a stuck chat let her avoid starting over. That is her account, not a guarantee; in practice, expect model choice to help most when a task stalls or needs a different style of reasoning, and treat it as a per-chat lever rather than a fixed setting.
What to weigh before relying on it
- Plan gating: The flexibility is described for Pro and Business, so confirm your plan includes it before building a workflow around switching.
- Per-chat, not global: The description points to choosing a model on a chat, so plan for consistency across a project rather than assuming one model governs everything.
- Real-repo context: Because each chat opens a real branch in a real GitHub repo and commits deploy automatically, changing models mid-project means reviewing the resulting commits like any other code change.
Practical next step
If you are comparing models for a specific job—say a data-heavy internal dashboard versus a marketing landing page—start one chat per candidate model on a template close to your target (the site lists dashboards, landing pages, components and apps), then compare the branches and deployed previews side by side. For broader context on how model-choice flexibility compares across AI builders, Vercel and GitHub document their own AI and deployment tooling.
How does Makedraft pricing compare across plans, and what does each tier include?
Makedraft's page does not publish plan prices or a full feature-by-feature tier breakdown, so a like-for-like comparison across tiers isn't possible from the site itself. What the page does confirm is that there are at least three paid tiers — Pro, Business, and an entry-level plan implied by the upgrade path — and that model choice is gated: Pro and Business are described as not locked to a single model, so selecting your own AI model is a paid-tier capability rather than something the base plan appears to include.
What the page indicates about tiers
| Area | What the page states | What it does not state |
|---|---|---|
| Model selection | Pro and Business can choose among AI models | Which models, or how many |
| Repo and deploy workflow | Every chat opens a real branch in a real GitHub repo; each commit auto-deploys | Whether limits differ by tier |
| Custom domain | Listed as a capability of the builder | Which tier unlocks it |
| Templates | Browsable library of apps, dashboards, landing pages, components | Whether any are tier-restricted |
| Price | A pricing page exists at /pricing | Any figures |
How to decide between tiers
Treat the model-selection line as the main functional divider the page reveals. If you only need the core loop — chat produces a real branch, commits deploy automatically, React and Next.js projects supported — the lower tier is the reasonable starting point, since the repo-and-deploy behavior is presented as the product's baseline rather than a premium add-on. Move up when switching models mid-project matters to you: the founder quoted on the page describes changing models on a chat that was stuck, which is a workflow benefit rather than a capacity one.
Two practical checks before choosing:
- Open Makedraft's pricing page directly for the actual numbers, since none appear in the product copy.
- Confirm whether custom domains and template access are tier-gated, as the page lists them as features without assigning them to a plan.
For a freelancer running client demos, the relevant question is whether the tier you pick keeps auto-deploy on every branch — the designer quoted on the page values exactly that, and it is the feature most worth verifying before committing to a cheaper plan.
What kinds of apps can I build with Makedraft's templates, such as dashboards or landing pages?
Makedraft’s template gallery is organized around practical project types rather than one narrow app category. Based on the site’s template listings, you can start from examples for landing pages, dashboards, components, apps and games, and image-generation playgrounds. The gallery also shows specific starting points such as a SaaS landing page, a photography-focused 3D gallery, an AI platform site, a creative studio page, and a map-transform tool.
What each template type is useful for
- Landing pages: Marketing sites for a product, service or launch. A SaaS landing-page template is the most direct fit if you need a polished page with sections for features, calls to action and signup.
- Dashboards: Internal tools or customer-facing interfaces that organize data and actions. Choose this when the main job is viewing, filtering or managing information rather than presenting a marketing message.
- Components: Smaller interface pieces you can adapt within a larger project. These suit teams that want a starting point for a particular interaction or layout rather than a complete site.
- Apps and games: Interactive projects where the template provides the initial structure. Check how much of the example is visual versus functional before assuming it matches your intended app.
- Image-generation playgrounds and creative tools: Interfaces for prompting, previewing or experimenting with generated images. These are relevant if your project centers on creative output rather than conventional forms and lists.
A useful distinction is between templates that primarily demonstrate a visual style and those that demonstrate a working workflow. A landing-page template mainly helps you arrange content and calls to action. A dashboard or playground template is more valuable if it already reflects the interaction pattern you need.
How to choose
Start by naming the first real task a visitor or team member must complete. For a client-facing product, that may be understanding an offer and signing up, which points toward a landing page. For an internal team, it may be reviewing records or changing settings, which points toward a dashboard. For a creative product, it may be entering a prompt and comparing results, which points toward a playground-style template.
Makedraft’s page evidence also says projects work with React and Next.js and that each chat opens a real branch in a GitHub repository, with commits deploying automatically. That matters when choosing a template: favor one whose structure you can inspect and modify in code, not just one that looks attractive in a preview. If you expect to connect authentication, data storage or payments later, check whether the template’s layout leaves clear places for those pieces.
For a concrete next step, open two or three templates in the category closest to your project and compare their page structure: navigation, primary action, data or content areas, and settings. Pick the one whose workflow you can describe in a sentence; visual restyling is usually easier than rebuilding the underlying interaction.
Do I need coding experience to use Makedraft, or can beginners ship a real project?
Beginners can ship a real project with Makedraft, but "no coding experience" and "no coding ability" aren't the same thing. Based on the site, Makedraft is an AI builder where each chat opens a real branch in a real GitHub repo, and every commit deploys automatically as its own live service. That means you're working directly with real code from the first message, not a throwaway preview you later have to export. The barrier to entry is lower than a traditional dev workflow, but you still benefit from being comfortable reading code, reviewing commits, and understanding what a repo and a deploy are.
What the site actually promises
- Real repo, real branch from message one — no sandbox or export step. You get a working codebase rather than a mockup.
- Automatic deploys per branch — approving a commit pushes it live.
- Model choice on paid tiers — Pro and Business plans let you pick the AI model, which matters when one model gets stuck.
- Custom domains — you can point your own domain instead of staying on a shared subdomain.
- React and Next.js projects — the stack is stated, so some familiarity with that ecosystem helps.
- Templates — landing pages, dashboards, components, apps and games are offered as starting points.
Where beginners will feel friction
The site's own testimonials hint at the real trade-offs. Priya K., an indie hacker, values that there's "nothing to migrate" because it was already her repo — great if you understand repos. Marcus T., a freelance designer, likes that "every commit I approve is live a minute later" — but notice the word approve: someone is reviewing changes. Sofia R., a founder, mentions switching models mid-chat when stuck, which assumes you can tell when output is wrong.
For a beginner, the concrete scenario is this: you pick a landing-page template, describe changes in chat, and get commits. You can ship something real. But when a deploy fails or a layout breaks, you'll be debugging React/Next.js code in GitHub. The AI reduces the writing; it doesn't remove the reviewing.
A practical decision rule
Ask yourself one question: will I read the code before approving a commit?
- If yes — a beginner with curiosity can absolutely ship. Start from a template, keep changes small, and review each commit before it deploys.
- If no — you can still get a site live, but you're trusting unreviewed code on a real domain. That's fine for a personal project, risky for anything customer-facing.
A sensible next step is to open a template closest to your goal and make one small change end-to-end — chat, commit, deploy, custom domain — before committing to a bigger build. If you want to compare approaches, official resources like GitHub for repo basics and Next.js for the framework will make the reviewing step much less intimidating.
Cloudflare
Namecheap Private Email
User reviews (0)