Bolt.new pricing is easier to evaluate when you treat AI tokens, hosting resources and team seats as separate constraints. Tokens buy interaction with the building agent; they are not a promise that an application will be completed, secure or deployed successfully. The same feature request may consume different amounts depending on the existing project and how many files need context.
As checked on October 26, 2026, Bolt lists a Free tier, Pro at $25 per month billed monthly, Teams at $30 per member per month billed monthly, and negotiated Enterprise terms. This article explains those advertised figures and shows a budgeting worksheet, rather than pretending to know a reader’s usage or the final taxes and annual discount at checkout.
Compare Free, Pro, Teams and Enterprise by constraint
Free provides public and private projects, up to 300,000 AI tokens per day and 1 million per month, according to Bolt’s current pricing page. It includes Bolt branding on websites, a 10 MB upload limit and limited hosted request capacity. Free is suitable for learning and a carefully scoped prototype, but the daily and monthly ceilings are both relevant: unused daily capacity does not remove the monthly cap. Verify details within your workspace because plan entitlements can change.
Pro is listed at $25 per month on monthly billing, with no daily AI-token limit and a starting allowance of 10 million tokens monthly. Pro also advertises custom domains, removal of Bolt branding, a 100 MB file-upload limit and expanded hosting and database capacity. Teams starts at $30 per member per month and adds centralized billing, administration and organizational collaboration. Enterprise uses customized terms; a sales conversation is not a public fixed price.
Separate AI tokens from other limits
The token counter measures interactions with Bolt’s agent. The vendor explicitly explains that synchronizing an application’s file system into the agent’s context can consume a large share of tokens; as a project grows, a similar-looking prompt may cost more. In contrast, file upload limits, website requests, hosting resources and database capacity are distinct constraints. A plan that has unused AI tokens can still face operational limits elsewhere.
Write a two-column inventory: “agent-building budget” and “running-app budget.” Put prompts, follow-up fixes and context size in the first column. Put traffic, databases, image uploads, custom domains, third-party services and backups in the second. Do not count a published website as free to run merely because the agent generated the source files.
Understand rollover and annual billing
Bolt says unused paid subscription tokens can roll over for one additional month, provided the paid subscription remains active. This is not indefinite savings and does not apply in the same way to every kind of allowance. If a project pauses for several months, accumulated tokens should not be treated like banked cash. Document whether your actual balance includes subscription tokens or other credits and the expiration conditions attached to each.
The pricing page advertises savings of up to 28 percent for annual billing but the headline $25 Pro and $30 Teams figures are monthly-billed prices. Never apply the maximum discount automatically to every tier. Record the exact annual charge, cancellation rules, required seats and tax in the checkout screen before comparing yearly and monthly options.
Make a token budget around milestones instead of prompts
Consider an illustrative small appointment-booking prototype. Phase one includes a static layout and navigation; phase two adds input validation; phase three adds login and persistence; phase four addresses testing and publish checks. Allocate a share of your monthly allowance to each phase, then reserve a separate contingency for revisions. These are work categories, not measured token-consumption predictions.
After the first real milestone, read Bolt’s token meter and capture what was delivered. If the first screen uses much more capacity than planned, simplify the scope or postpone database features. A budget should respond to observed usage rather than an invented average number of tokens per button or screen. A single request for a repo-wide rewrite may be more expensive than several narrowly scoped changes.
When the Pro tier is worth paying for
Pro makes sense when you regularly exhaust Free’s daily or monthly AI quota, genuinely need unbranded publishing, larger uploads, domains or expanded resources. It is less persuasive if you mainly need a one-time static mock-up and can export or maintain conventional code without continued agent editing. Define your required deliverable first: a working clickable demo is a different commitment from a production system with authentication and paid users.
Before upgrading, review the preceding week of prompts and mark each outcome as accepted, revised or abandoned. If frequent retries come from vague acceptance criteria, buying ten times more tokens will not automatically solve the problem. Specify data shapes, edge cases and success conditions, and put high-risk integrations behind a manual review gate.
Price a team by people, not by a shared token pool
Bolt says Teams is priced per member and that each paid teammate receives their own monthly token allotment rather than everyone drawing from one common token bucket. A three-person team on a $30 monthly per-member headline therefore begins with an illustrative $90 monthly recurring seat total, before taxes or any additional service charges. This multiplication is arithmetic, not a personalized invoice or a promise about usage.
Ask who edits code, who reviews changes, who can publish and who controls billing. If only one person builds while three people inspect screenshots, not all three necessarily need paid editing seats; verify collaboration entitlements and organizational policies. Teams should be chosen because administration and separate productive capacity matter, not simply because the organization has multiple employees.
Do not confuse hosting allowances with reliability
The published comparison lists included website hosting, maximum web-request allowances and database access, with higher limits on paid plans. Those figures describe capacity policy, not uptime guarantees, successful payments or secure user authentication. High-traffic launches also require backups, alerting and an owner who can recover from a faulty deployment. External email, payment and AI APIs may bill separately from Bolt.
Run a small end-to-end test on a staging address. Verify failed login, empty form inputs, unauthorized record access, slow responses and rollback to a known version. If the app handles personal information or customer payments, a real security review is mandatory regardless of the builder tier. Keep production secrets outside prompts and source-control history.
Use a purchase checklist and a stop rule
Before choosing a plan, list the expected collaborators, the smallest acceptance-tested prototype, the desired custom domain, traffic assumptions, sensitive data categories and third-party services. In your account, inspect the current token meter and billing dashboard. Keep a clearly stated maximum monthly spend and define who can approve an upgrade or top-up.
If the actual agent usage, build quality or infrastructure bill differs sharply from expectations, pause new features and review the last accepted version. Upgrade only when a documented limit is blocking a valuable outcome. Recheck official pricing and billing terms at purchase time: this is a dated editorial snapshot and not a contractual quote.
How We Research
We reviewed the current primary-source documentation as of October 26, 2026: Bolt official pricing; Bolt token and billing help; Bolt planning and prompting documentation. We distinguished documented plan features and limits from editorial workflow suggestions, and we compared the intended reader task with the previously published LabFin material. We did not purchase plans, perform product benchmarks, run code or verify an individual account’s entitlements. Where product pages conflict or terms may change, we have identified the uncertainty and recommend confirming checkout or live account details.
Frequently Asked Questions
The current public Bolt pricing page lists 300,000 per day and 1 million per month. Both ceilings apply; check the live plan for changes.
The advertised $25 figure is for monthly billing. Annual offers, discounts, taxes and account-specific terms may differ.
No. AI tokens and hosting or database capacity are separate usage concepts, each with its own plan limitations.
Bolt says each paid Teams member receives an individual monthly token allotment; it is not a single pooled balance.
No. Bolt describes paid subscription-token rollover for one additional month while the relevant paid subscription remains active.