Bubble for Startups: MVP Costs, Partner Credits and Lock-In

Bubble (bubble.io) can carry a startup from first prototype to a live, paying product, provided the founders accept two trade-offs early: the app never leaves Bubble as code, and the bill follows server work rather than headcount. Building costs nothing on the Free plan. Launching does not. Bubble's Starter plan (Web) is $32 per month billed annually or $39 per month billed monthly, per app, and it is the first tier that can put a web app live on your own domain ^[1]. Two cofounders who both edit need Bubble's Growth plan (Web), at $119 per month billed annually or $134 per month billed monthly, per app. The fit is strongest for data-heavy SaaS products and marketplaces.

From prompt to a clickable product, before any invoice

The cheapest phase is the one before launch. On Free, a project gets 50K workload units (WU) and 500 AI credits a month, one app editor, 200 database things and 0.5GB of file storage, with Bubble branding and only a development version. That is plenty to model the data, wire the workflows and click through the product with the founding team. Nothing on Free goes live, though: no live web version, no custom domain, no App Store or Google Play release. Schedule the upgrade before the first customer demo; our guide to Bubble's Free plan lists every Free limit.

Bubble AI shortens the blank-page stage. App generation turns a prompt into pages, a database with sample data, workflows, privacy rules and sign-up flows in roughly 10 to 15 minutes, by Bubble's September 2026 account, and it spends no AI credits. Follow-up edits through the Agent do spend them. Agent 2 left beta in October 2026; it can show a plan to approve before changing the app, in three modes (Ask confirms every change, Auto is the default, Approve applies without waiting), and edits pages, data types and workflows. The manual still lists gaps, such as custom events and payment actions in backend workflows, which a founder wires by hand in the visual editor.

Workload units matter even before launch. They meter server work (searches, writes, API calls, scheduled workflows, emails), and the Development version draws on the same monthly allowance as Live. Bubble's manual prices a page load at 0.15 WU and each thing written at 0.5 WU, while client-side actions such as showing a group cost none. One careless loop can hurt. A recursive workflow without a stop condition can drain a month's allowance, and the Logs tab shows which workflow spent it.

The spending ladder: Free, then Starter, then Growth

Growth earns its price for three startup reasons. It adds a second app editor for a cofounder or first hire, premium version control with up to 10 custom branches so a feature can be built away from Main, and unlimited advanced security tests plus two-factor authentication for the app's own users. Starter offers five basic security tests and flags advanced issues without details. A solo founder can stay on Starter for a long time. A two-person team where both people build hits the editor cap on day one.

Every plan is priced per project, so a customer app and a separate admin app need two plans. A project that pairs web and native mobile shares one plan, one database and one WU pool, on the Web & Mobile tab, where Bubble's Starter plan (Web & Mobile) is $59 per month billed annually or $69 per month billed monthly, per app. Store accounts come on top. Apple charges $99 a year for its developer program and Google Play $25 one time, both paid to those companies. Bubble labels its native mobile builder beta as of November 2026.

Traffic is the second trigger. Past the plan's allowance, overages bill at $0.30 per 1K WU, on by default and invoiced separately each month. Switching them off takes the app offline at the limit until overages are switched back on, a tier is bought or the next billing period starts, a poor surprise during launch week. Steady extra demand is cheaper as a workload tier: Tier 1 adds 200K WU at $26 per month on an annual plan or $29 per month on a monthly one. By our arithmetic, 100K extra WU at the default rate costs $30, and Tier 1 wins once the monthly extra passes about 87K WU on an annual tier or about 97K WU on a monthly tier.

Annual billing locks the year in. The toggle advertises savings of up to 20%, yet by our arithmetic Bubble's Starter plan (Web) costs $384 a year billed annually against $468 billed monthly, a saving of $84, about 18%. An annual plan renews for a year unless canceled at least 30 days before the anniversary, early cancellation does not end it, and subscription charges are non-refundable ^[2]. A pre-revenue team still testing demand is usually better served by monthly billing; the Bubble pricing guide covers the larger tiers and the Mobile tab.

Startup credits and Launch Lab, read closely

Bubble said in December 2025 that it has no fund investing in Bubble startups, so the money on offer is billing credit. Several accelerator and perk programs hand out Bubble credits, which are separate from the AI credits inside each plan and are spent against the subscription itself. Bubble's partnership pages list the amounts below, and each program has its own membership test, so check eligibility with the program before planning a launch budget around it:

• On Deck, AWS Activate and Harvard Innovation Labs members: $2,500 in Bubble credits, valid for six months.
• Microsoft for Startups members: $3,000 in Bubble credits for six months, or $3,500 for founders from underrepresented minority backgrounds.
• Startup Grind: up to $500 for community members, or $100 for new users through a second offer, each valid for three months.
• Stripe Atlas startups: $500 in Bubble credits for three months, plus $50 off Atlas incorporation.

The fine print narrows the offers. Every listed program limits credits to monthly subscription plans, and the AWS Activate page adds that credits go to new paying Bubble users only, exclude Dedicated plans and one-time plugin or template purchases, and work once per organization ^[3]. That page also asks for the same email as the program membership plus a linked card, and the credit cannot be stacked with another coupon. The monthly-only rule matters. A founder spending partner credits pays monthly rates, so the annual saving waits until the credits are used up.

Launch Lab is the other free program. Bubble opened it in January 2026; it is application-based and aimed at apps launching within one to three months, or launched within the past month. Over about two to three weeks, Bubble produces storytelling, video and amplification across its own channels, and the founders keep the content for fundraising decks or recruiting. It is marketing, not paid user acquisition. An earlier Q1 2026 Launch Program gave sign-ups $100 in Bubble credits; that cohort ran from January to March 2026 and belongs to history.

Founders who shipped, in Bubble's telling

Bubble's customer pages supply the founder stories, and every figure in them comes from Bubble or the customer. SuiteOp, a hospitality startup, built its native app in two months against an estimated six to eight, Bubble said in August 2025, and Bubble puts its audience at 30,000 daily users. BuyTicket, a ticket-resale marketplace, was built in two months by co-founders still in high school and, Bubble said in August 2025, reached 20,000 users in its first week. Ohana, a subletting marketplace, handled 2,700 sublease transactions over 14 months, Bubble said in July 2025.

Bubble said in July 2026 that My AskAI serves more than 40,000 businesses. A broader signal comes from Bubble's study of more than 250,000 apps as of mid-August 2026: Bubble says 70% of monetizing apps were built by a solo founder, and that the median gap from deploy to the first payment trigger was four days ^[4]. Read these as the vendor's chosen winners, not base rates. Nothing in them shows how many startups stalled, and the study counts a payment workflow, not revenue that lasted.

Payments run through the Bubble-built Stripe plugin, since Bubble does not charge cards itself. For multi-tenant SaaS, the manual recommends a parent Company data type tying every record to its tenant, enforced with privacy rules. Native mobile apps can sell subscriptions through Apple and Google billing but not one-time digital purchases, which rules out some consumer business models. Check that constraint before promising a mobile paywall to investors.

Lock-in questions to settle before a funding round

An investor or technical hire will ask what happens if the company outgrows Bubble. The honest answer: there is no code export. The app runs only on Bubble, so leaving means rebuilding the logic elsewhere. What travels is the data, exportable as CSV, JSON or NDJSON, plus a raw JSON export of application logic; full app configuration export needs Growth. Through the Bubble MCP (beta), an outside AI agent can recreate the app in code, without guaranteed fidelity, after which you supply hosting, database and auth yourself. Lovable generates source code instead, as our Lovable review explains.

Ownership is cleaner than the lock-in suggests. The Terms give you your content, design, workflows and data, and let you reuse workflow and design on other platforms; Bubble keeps the underlying engine. Bubble also states it would release its source code under an open-source license if the business shut down, a pledge rather than an escrow. Diligence will probe the plan gates too. Self-serve apps are hosted in the US with no region choice and no uptime SLA, while single sign-on for Bubble accounts, dedicated instances and region choice sit on Enterprise, priced by quote.

Security reviews will reach the October 2026 incident. Bubble's preliminary root-cause analysis (November 2026) says an attacker in October 2026 got partial access to Bubble's production systems and some user data, including private plugin data such as credentials plugin developers had saved. It also targeted individual apps through platform issues that reached stored credentials or data even in correctly set-up apps, and through settings such as privacy rules that granted access; affected owners were contacted. Bubble says its review had not shown app databases being read through the internal access, and a final RCA will follow. Bubble says its platform is SOC 2 Type II compliant, yet that status does not extend to your app. Privacy rules stay the founders' job.

The bottom line

Bubble suits founders who want a working, data-heavy product without hiring engineers first, and who accept that the product will live on Bubble's platform for the foreseeable future. Build on Free, move to Bubble's Starter plan (Web) when real customers arrive, and step up to Growth once a second person builds; if only the WU curve climbs, add a workload tier instead of changing plan. Claim partner credits on a monthly plan, watch overages from launch week, and settle the no-code-export question openly with investors before a round, not during one.