Base44

What Base44 is

Base44 turns a plain-language description into a working web app, and unlike most of the vibe-coding cluster it hands you the whole stack rather than the front end. Database, authentication, file storage, auto-generated API endpoints, payments through a Stripe connection, email, hosting and a custom domain are all in the box. The competitive claim is not that it writes better React than Lovable or Bolt.new; it is that you do not go and wire up Supabase, Clerk and Vercel afterwards.

Wix bought it in June 2025, roughly six months after founding, for about $80 million up front plus earn-out payments running through 2029 and around $25 million in retention bonuses for Base44 staff. Wix said at the time that Base44 would continue to operate as a distinct product and business, and that has held: separate brand, separate pricing, separate docs, with a Wix copyright line in the footer as the only visible tell. The earn-outs have kept triggering, which means the revenue targets were beaten rather than missed. Wix disclosed a further payout in its second-quarter 2026 results, and press coverage of its filings puts Base44's annual recurring revenue run rate at roughly $150 million as of May 2026.

In June 2026 Base44 announced Base1, a model it presents as its own. The precise version is that Base1 is fine-tuned on top of an existing open-source foundation model rather than trained from scratch, using interaction data from building sessions on the platform, and TechCrunch's account of the launch is that it does not yet consistently beat the strongest frontier models. It sits in the model picker alongside third-party models rather than replacing them, and the picker itself is a Builder-and-up feature. We have not benchmarked it and Base44 has published no figures we could check.

The thing to understand before you buy is the metering, because it is unusual and it is where the money goes. Base44 runs two balances. Message credits are spent when you talk to the builder AI. Integration credits are spent when your published app does something at runtime: sends an email, generates an image, runs an LLM call, answers a question through an in-app agent, fires an automation. That second meter is the one people do not price in. Your building and your users' activity draw from the same monthly allowance, it does not roll over, and when the integration balance hits zero your live app starts failing for the people using it.

Pricing

Base44 publishes its prices, its credit allowances and, unusually for this category, an itemised estimate of what each action costs in credits. Plans are priced per workspace, not per seat: there is no per-user charge, and invited members draw from the workspace's shared pool. The pricing page opens on the Yearly toggle, so the first numbers you see are twelve-month commitments displayed as a monthly figure. The Monthly column is one click away and costs exactly 20 percent more at every rung.

All figures in US dollars, per workspace, read from base44.com/pricing in September 2026 with both billing toggles driven live, and cross-checked against Base44's own July 2026 blog post on plan costs, which states the same monthly and annual figures in plain text.

Elite is a floor. The public pricing page shows a flat $160 billed annually or $200 monthly, with no dropdown. Base44's billing documentation says the Elite tier's message and integration credit volume is chosen from an in-app dropdown, so higher rungs exist at prices that are not published anywhere we could read without a paid login.

Workspaces multiply the bill. Each workspace has its own plan and its own shared pool of credits. A member who wants to build in their own workspace needs a separate paid plan for it, and you can own up to four workspaces including the personal one created at signup. Two teams on two workspaces is two subscriptions.

Discounts. Students, teachers and academic staff get 50 percent off Starter or Builder after verifying a school email. Non-profit pricing is case by case.

Yearly terms. Base44's Terms of Use make yearly fees non-cancelable and non-refundable, and its billing docs say you cannot switch an active yearly plan back to monthly yourself; support reviews requests case by case. The pricing FAQ's line that you can cancel any time is true only in the narrow sense that changes take effect at the next billing cycle. Cancelling drops the workspace to Free at period end, which caps you at 5 apps; apps stay live but paid features go away. Subscriptions bought through the iOS App Store are billed and refunded by Apple, and you cannot hold a paid plan on both platforms at once.

Top-ups. Base44 says twice that you can top up credits at any time, once in the pricing FAQ and once on its blog. Neither surface says what a top-up costs or how to buy one, and the credits documentation's own list of ways to get more credits does not include it. That list is: upgrade your plan, wait for the monthly reset, redeem a coupon code, claim a one-time 20 credit reward for an approved social post, or use the referral program.

How to budget credits correctly

Base44's docs publish rough credit costs and label them estimates based on standard models and average complexity, subject to change at Base44's discretion. Manual visual edits, dragging, layout and direct text changes, cost nothing. These are the figures to budget from.

Worked through, Starter's 2,000 integration credits is about 2,000 plain emails, or about six and a half minutes of generated video, or roughly 66 in-app agent messages if the agent runs on an Opus-class model. That is the number to budget from, not the 100 message credits in the headline.

Two constraints make this harder than it looks. Base44's docs state there is no way to preview an integration credit cost before running an action; the recommended method is to run it once in a controlled test and read the execution logs afterwards. And unused credits of both kinds expire at the end of each monthly cycle with no rollover, so a quiet month banks nothing for a busy one. Per-app and per-day usage dashboards, plus a per-message credit readout in the builder, mean overspend is at least diagnosable after the fact.

What we checked, and what we did not

Reviews of platforms this wide often imply a score covers everything. This one does not, so here is the boundary in plain terms.

Checked: every price and credit allowance on the pricing page under both billing toggles; the plan highlights per tier; Base44's blog post on plan costs; the credits documentation for per-action costs, the free tier's daily cap, the rollover rule, the no-preview rule and the out-of-credits behaviour; the billing documentation for workspace, Elite, academic, yearly and App Store rules; the GitHub sync documentation for the tier gate and the branch requirement; the security page for certifications; and Wix's acquisition announcement and 2026 earnings disclosures.

Not checked: anything that requires an account. We did not build an app, so we cannot speak to build quality, generation speed, error rates, how closely the published credit estimates match real spend, or whether the app-breaks-at-zero failure is as abrupt in practice as the documentation describes. Base1 is unbenchmarked. Superagents' plan requirement and beta or general availability status are not stated by the vendor. HIPAA is not mentioned on the security page either way. We found no published penetration test following the July 2025 authentication bypass.

That means the score below is a reading of what Base44 publishes about itself, checked for internal consistency, and it gives no credit in either direction for what we did not run.

What each tier actually unlocks

The pricing page's plan highlights are cumulative, and the differences between tiers are about capability as much as credits. This is the table to read before choosing, because two of the rows are the ones the marketing copy glosses over.

The row that matters is GitHub. The pricing page's FAQ answers the ownership question with an unqualified yes and says two-way GitHub sync exports the full source to your own repo "at any time", with nothing locked to the platform. The same page's plan highlights list GitHub integration under Builder, Pro and Elite only, and Base44's blog says outright that GitHub integration starts at the Builder tier. On Free and Starter the portability the FAQ promises is not available to you. That is $40 a month billed annually, $480 up front, or $50 month to month, before the code is yours to take.

Where Base44 is the strongest pick

Base44 is at its strongest when the backend is most of the work. An internal dashboard, a client portal with roles, a back-office system that reads and writes Salesforce or Snowflake, an MVP that needs real login on day one: these are jobs where the front-end-first builders leave you assembling a database, an auth provider and a host afterwards, and Base44 does not. Managed connectors for Google Workspace, Slack, Salesforce, GitHub, Snowflake and Databricks, plus custom MCP server connections from the builder chat, mean the integration work that usually eats week two is largely a prompt.

The second argument is that you control the traffic. Internal tools have a known number of users doing a known number of things, which is exactly the situation where a per-action runtime meter is predictable rather than dangerous. The same meter pointed at an open consumer signup page is a different proposition, and we cover that below.

Best for

Founders, operators and internal-tools teams who want a working app with a real backend rather than a front end they then have to staff, and who are comfortable budgeting from a per-action credit table instead of a flat price. It suits internal dashboards, back-office systems, customer portals and MVPs where you control the traffic and can therefore predict the runtime credit burn. Builder at $40/mo billed annually, or $50 monthly, is the honest starting tier, because it is the first one with source-code export and model selection. Not for anyone shipping a consumer app with unpredictable volume, where a traffic spike converts directly into a broken feature for your users, and not for teams who need a fixed monthly cost they can put in a budget line.

Key features

• Full backend included: managed database, user authentication, file storage and auto-generated API endpoints, with no external service to wire up
• Built-in hosting and one-click publish, with custom domain connection from Starter at $16/mo billed annually or $20/mo monthly
• GitHub two-way sync that keeps the full source in a repository you own, from Builder upward; the sync branch must be named main
• AI model selection from Builder upward, including Base44's Base1 model alongside third-party frontier models
• Superagents: autonomous agents that connect to Gmail, Calendar, Drive, Slack, Discord, LinkedIn, Notion, HubSpot, Telegram and Wix, and reach you on WhatsApp, Telegram, Slack or the browser
• Managed connectors for Google Workspace, Slack, Salesforce, GitHub, Snowflake and Databricks, plus custom MCP server connections in the builder chat
• In-app AI agents and automations (scheduled, data-event, agent-triggered or connector-triggered) that run inside your published app
• Discuss mode for planning without touching the live app, at about 0.3 message credits per message versus roughly 1 to 4 in Build mode
• Per-app and per-day credit usage dashboards showing which app and which source burned what
• Backend functions and a base44 CLI plus SDK for local development, from Builder upward
• SOC 2 Type II and ISO 27001 certification on all tiers, GDPR with a Data Processing Agreement on request, TLS 1.2+ in transit and AES-256 at rest
• Enterprise controls: enforced organizational SSO, audit logs with a streaming API, IP allowlisting, per-member credit limits and data version history

Best-fit use cases

• Internal back-office tools and admin dashboards where the database and auth are most of the work and traffic is predictable
• Customer portals and client-facing web apps that need real login, roles and a database on day one
• MVPs and early-stage products where shipping something live beats owning the infrastructure
• Replacing a spreadsheet-plus-Zapier workflow with an actual app that has automations and scheduled jobs
• Operations apps that need to read and write Google Workspace, Slack, Salesforce, Snowflake or Databricks through managed connectors
• Building an app to a working state on Base44, then exporting via GitHub two-way sync and continuing in a normal dev environment
• Autonomous agent workflows via Superagents, where an agent acts on your tools and reaches you on WhatsApp, Telegram or Slack
• Student and academic projects, at 50 percent off Starter or Builder

Pros

• The stack really is complete. Database, auth, storage, payments, email, hosting and domain in one place is a materially different offer from tools that hand you a front end and a shrug
• Credit costs are published in detail, per action, with numbers. Most of this category publishes nothing comparable
• SOC 2 Type II and ISO 27001 apply to all tiers, not just Enterprise, which is unusual at a $16/mo entry price
• Genuine market traction: a roughly $150 million ARR run rate as of May 2026 and repeated earn-out payouts from Wix, so the platform is not going anywhere in the near term
• Wix ownership without Wix absorption. Separate brand, separate pricing, separate docs, and the product has kept shipping
• Real source-code portability via GitHub two-way sync, so the app is not trapped, provided you are on Builder or above
• Per-app and per-source credit dashboards, plus a per-message credit readout, so overspend is at least diagnosable after the fact
• On paid plans, automatic AI error fixes are free, so the AI cleaning up its own mess does not bill you
• Manual visual edits, dragging, layout and direct text changes, cost nothing at all
• The free plan is a real free plan: core features, auth and database included, not a crippled preview

Cons

• Integration credits meter your live app, not just your building, so a successful launch eats the same balance that builds the product
• When integration credits run out, your app's users hit a generic error with no explanation, per Base44's own documentation
• No credit rollover. Unused message and integration credits expire at the end of every cycle
• No way to preview an integration credit cost before running an action. The documented workaround is to run it once and read the logs
• GitHub two-way sync, the mechanism behind the pricing FAQ's claim that nothing is locked to the platform, requires Builder at $40/mo billed annually ($480 up front) or $50/mo monthly, and the FAQ does not say so
• The pricing page defaults to the Yearly toggle, and yearly fees are non-cancelable and non-refundable under the Terms of Use
• The free plan is throttled to 5 message credits a day, so its 25 monthly credits cannot be spent in one sitting
• On the free plan, the AI's own automatic error fixes consume your message credits; that only becomes free once you pay
• Credits are non-refundable for AI mistakes. A prompt that breaks your app still costs you, and so does every prompt that fixes it
• Published credit costs are labelled estimates and are subject to change at Base44's discretion
• Elite's $160/mo is the entry rung of a credit band chosen from an in-app dropdown, so the top self-serve tier's real cost is not on the pricing page
• Top-ups are promised in two places and priced in none
• A critical authentication bypass was disclosed in July 2025 in the auth layer Base44 sells as a reason not to build your own

What to watch out for

Five things, in order of how much they will cost you.

Your live app runs on the same meter you build with. Integration credits are consumed at runtime by your published app: every signup email, image generation, LLM call, agent reply and automation run. Base44's docs are unusually blunt about what happens at zero. Any action requiring integration credits fails, and the person using your app sees "a generic error message without any reference to integration credits". The documentation includes a screenshot of it. Read that twice. It means a good launch week can take your signup emails offline mid-month, and your user sees an unexplained failure with no path to resolution. Nobody in your funnel knows the cause. Neither will you until you check the dashboard or the email Base44 sends.

You cannot see a cost before you incur it, and nothing carries over. The docs state there is no way to preview an integration credit cost before running an action, and recommend running it in a controlled test and reading the execution logs afterwards. Unused credits expire monthly. The published costs are explicitly estimates, subject to change at Base44's discretion. So the meter is retrospective, the allowance is use-it-or-lose-it, and the rates are not contractual.

"You own your code" carries a price tag the FAQ does not mention. Base44's pricing FAQ says two-way GitHub sync exports the full source to your own repo at any time and that nothing is locked to the platform. On the same page, GitHub integration appears in the plan highlights for Builder, Pro and Elite only, and the documentation and Base44's own blog both put two-way sync at Builder or higher. That is $480 a year prepaid, or $50 a month. On Free or Starter, the portability the marketing page promises is not available to you. If code ownership is why you are choosing Base44, budget for Builder from day one, not Starter.

The default toggle costs you the right to change your mind. The pricing page loads on Yearly, so the number you see is a twelve-month prepaid commitment displayed as a monthly figure. Base44's Terms of Use make yearly fees non-cancelable and non-refundable, and the docs confirm you cannot switch an active yearly plan back to monthly yourself; support considers exceptions case by case. Annual also buys entitlements, not only a discount: the free domain for a year and the 25 shareable credits on Builder and above disappear from the plan highlights the moment you click Monthly. If you are not certain, click Monthly and pay the 20 percent.

The auth layer failed once, in public. On 9 July 2025, shortly after the Wix acquisition, Wiz Research disclosed a critical authentication bypass: undocumented registration and email-verification endpoints let anyone create a verified account on any private Base44 app using only the app_id, which is visible in the app's own URL and manifest. That defeated SSO. Base44 fixed it in about 24 hours and there is no evidence it was exploited. We are not holding a patched 2025 bug against the 2026 product, and the disclosure handling was good. But the thing that broke was the built-in authentication, which is precisely the component Base44 asks you to trust instead of building your own. If you are putting HR data or customer records behind it, that is a reason to ask for the current pen-test report rather than to take the SOC 2 badge at face value.

Is Base44 worth it?

Yes, for the buyer it is actually built for. If the job is an internal tool, a portal or a back-office system, the difference between a front end and a working app is a database, an auth provider, a host, an email service and a payments integration, and Base44 ships all of them behind one prompt. That is the most complete stack in this cluster, the company behind it has disclosed financials rather than a funding announcement, and the security paperwork applies at $16 a month rather than only at Enterprise.

It scores 4.2 rather than higher for one reason, and it is not the product. The pricing model routes your users' activity through the same expiring balance you build with, gives you no way to see a cost before you incur it, publishes its rates as revisable estimates, and documents that a drained balance fails your live app in front of the people using it with no explanation. That is a structural risk to the thing you built, not a billing annoyance. The gating of code export to Builder, on a page whose FAQ says export is available at any time, costs a little more.

That puts it at the bottom of its own comparison set on this site: level with Softr at 4.2, a notch below Bolt.new and Replit at 4.3, below Lovable and v0 at 4.4, and well below Emergent at 4.7. It would move up on a clearer credit model, a cost preview before an action runs, or any mechanism that keeps a live app serving when the balance hits zero. It would also move up on hands-on testing: we did not build anything on it, and a page written from documentation should not claim more than 4.2.

How we arrived at 4.2

The score is weighted across ten categories. Because nothing here was tested hands-on, every line is a reading of Base44's own published material and the company's disclosures, checked against each other.

The 5 point score is a product rating rather than a linear map of that total, and it reflects a genuinely complete platform with real market traction, held back by a metering model that turns your users' success into your app's failure mode.

FAQ

Pricing, plan limits and feature claims verified September 2026 from base44.com/pricing under both the Yearly and Monthly toggles, Base44's own blog post on plan costs, its credits, billing and GitHub documentation, and its security page. Company facts come from Wix's June 2025 acquisition announcement and press coverage of its 2026 quarterly disclosures. No account was created and no app was built for this review. ToolChase received no payment for this review. Scores, rankings and list positions are never sold. See our methodology for how we score tools.