Bolt.new Review 2026

Bolt.new is worth a shortlist when the pain is slow setup: you want to describe an app, see files appear, preview instantly in the browser, and publish on Bolt Cloud without wiring Netlify, Supabase, and auth yourself. StackBlitz built its reputation on in-browser WebContainers; Bolt.new layers an agentic chat builder on top so non-engineers can ship landing pages, internal tools, and MVPs. Official positioning includes import from Figma and GitHub, AI image editing on Pro, NPM registry support on Teams, and Enterprise SSO with audit logs for regulated buyers.

The buying mistake is treating 10,000,000 Pro tokens as ten million chat messages. Bolt.new syncs project files into the model context, and token usage scales with the size of the filesystem attached to each turn. A greenfield todo app and a imported Create React App with dependencies, config files, and assets can consume wildly different token budgets on the same natural-language instruction. Free adds a daily cap of 300,000 tokens on top of the 1,000,000 monthly pool, which is enough to learn the product but not to run a team pilot on a real repo.

How we evaluated this

This is a desk review, not a multi-week production benchmark. We checked the official Bolt.new homepage, pricing page, and support documentation on token usage on August 20, 2026. We cross-read the AI app builders buyer guide on this site and prior notes that compared Bolt.new to Lovable and Replit on portability and hosting. We did not independently audit Enterprise security controls, run a paid load test against Bolt Cloud request limits, or measure exact token counts on a proprietary customer repo.

We weight prompt-to-preview latency, token economics on small versus large file trees, Bolt Cloud hosting and database claims, import fidelity from GitHub and Figma, Teams billing quirks, and exit paths when generated code must live elsewhere. Bolt.new markets unlimited databases on Free and paid tiers; treat database and request limits as operational caps you validate in a pilot, not as unlimited production scale.

Cost model

Token burn grows with repo size because filesystem sync dominates each turn

Bolt.new looks affordable because Pro starts at $25 per month with 10,000,000 tokens and no daily limit. The unit that actually constrains a maturing project is how much of your file tree the agent re-syncs into context on every instruction. That is a different shape from a chat assistant where only the messages count, or from a seat-based IDE like Cursor where usage pools attach to the editor rather than to synced project bytes.

• Filesystem sync is the hidden multiplier. Official documentation states token usage scales with project filesystem size synced to the AI. A five-file landing page and a forty-file React app with lockfiles, env samples, and public assets are not the same meter, even if your prompts are identical.
• 10M Pro tokens depletes faster on large imports. Import a mature GitHub repo or let a week of iterative edits inflate the tree, and the same "fix the navbar" request costs more tokens than it did on day one. Greenfield MVPs feel generous; inherited codebases feel expensive.
• Free has two caps. 300,000 tokens per day and 1,000,000 tokens per month, plus Bolt branding on published sites. That is a learning tier, not a team production plan.
• Pro rollover is conditional. Unused tokens roll forward one extra month while an active subscription remains. Pause billing and rollover rules change; confirm at checkout.
• Teams tokens do not pool. At $30 per member per month, each seat gets its own token allocation. Centralized billing helps finance, but a heavy builder and a light reviewer do not share one bucket.
• Cloud limits are separate meters. Free includes 333,000 web requests; Pro includes 1,000,000 web requests per month. Traffic spikes on a successful launch can hit request caps even when tokens remain.

Worked example on public USD rates checked August 20, 2026. A solo founder on Pro at $25 per month starts with 10,000,000 tokens. Week one on a greenfield SaaS shell might spend a few hundred thousand tokens across fifty chat turns. Week four, after GitHub import added components, tests, and config, the same fifty turns can burn several times more because each sync sends a larger snapshot. A three-person product team on Teams at $90 per month total does not get a shared 30M token pool; each member meters separately, so the designer and the engineer need individual usage discipline.

If you expect to live inside an imported monorepo, model Pro as a pilot tier and talk to sales about Enterprise before you commit marketing spend to a Bolt-hosted URL. Compare the same workflow on Lovable and Replit with a file tree matched to your production shape, not a hello-world demo.

Who Bolt.new fits best

Surfaces to verify

Bolt.new surfaces are the chat builder, Bolt Cloud hosting, databases, auth, GitHub and Figma import, and Teams administration. They are not a customer-support inbox or a CRM. Map each surface against your launch checklist before you standardize on Bolt for client work.

• Chat builder: confirm how edits apply across multi-file projects, whether diffs are reviewable before accept, and how rollback works when a prompt breaks the build.
• Preview and WebContainer runtime: test your framework, native modules, and env var patterns. Not every npm package behaves identically to local Node.
• Bolt Cloud hosting: validate deploy URLs, custom domains on Pro, SSL, and what happens when you exceed 333,000 Free or 1,000,000 Pro web requests.
• Databases: official copy advertises unlimited databases even on Free. Confirm schema tooling, backups, connection strings, and whether production data residency meets your policy.
• Auth: test sign-up flows you plan to ship. Generated auth is fast until you need SAML, fine-grained roles, or audit trails.
• SEO and publishing: on Pro, verify meta tags, sitemap behavior, and AI image editing output quality for marketing pages.
• GitHub import: import a repo representative of your stack. Measure token burn on the first ten edits, not only import success.
• Figma import: confirm layout fidelity, responsive behavior, and how much cleanup chat turns require after import.
• Teams workspace: validate centralized billing, per-member token isolation, private NPM registries, and design system prompts with two seats during pilot.
• Upload limits: 10MB on Free versus 100MB on Pro affects asset-heavy apps and document portals.

Build workflow and handoff

The intended loop is: describe the product in chat or import design or code, let Bolt generate and modify files inside the WebContainer, preview instantly, iterate with follow-up prompts, then publish on Bolt Cloud with optional custom domain and SEO on Pro. The buying question is whether that loop survives your file tree size, your compliance rules, and the moment you need git discipline, staging environments, or non-Bolt hosting.

Founder-led teams often start greenfield: landing page, waitlist, simple CRUD app with Bolt-managed database and auth. Agency-style teams may start from Figma for visual fidelity, then spend chat turns wiring backend behavior. Engineering-adjacent teams may GitHub-import an existing starter and discover token burn dominates once the repo grows.

Handoff quality is the difference between a demo URL and an operable product. Test whether you can export or sync to GitHub cleanly, what secrets management looks like, how database migrations behave across deploys, and whether your team can continue development in Cursor or Replit without rewriting half the scaffold. Also test failure recovery: when a prompt breaks routing or auth, how many tokens does recovery take on a large tree?

Pricing checked August 20, 2026

Bolt.new pricing: tokens, Cloud limits, and Teams isolation

Public USD packaging is Free, Pro, Teams, and Enterprise. Rates below are from bolt.new/pricing, checked August 20, 2026. Pro displays $25 per month billed monthly with roughly 28% savings on annual billing per the pricing page toggle.

Source: bolt.new/pricing and token usage documentation, checked August 20, 2026. Confirm checkout price, annual discount, rollover rules, Teams per-seat token allotments, and Enterprise scope before purchase. Token usage scales with synced filesystem size per official guidance.

AI capability: chat building, imports, and image editing

Bolt.new centers on an agentic chat that creates and edits project files, runs terminal commands inside WebContainers, and redeploys previews. The AI is not a sidebar autocomplete; it is the primary author of scaffolding, components, API routes, and database schema for many users. That makes prompt quality and project scope management core skills, not nice-to-haves.

Import flows matter as much as greenfield generation. Figma import targets teams with approved visual design who want code without manual recreation. GitHub import targets teams with starter repos or open-source bases. Both paths increase filesystem size and therefore token burn on every subsequent turn, which is why import success alone is a weak pilot signal.

Pro adds AI image editing for marketing assets and page visuals. Useful for landing pages when you do not have a separate design tool chain. It does not replace a dedicated design system unless Teams design-system prompts cover your brand rules. Compare UI-focused generation on v0 if the deliverable is component code inside an existing Vercel workflow rather than a full hosted app on Bolt Cloud.

Feature areas to verify in a pilot

• Greenfield app from a ten-sentence brief: auth, database, and deploy on Bolt Cloud.
• GitHub import of a repo with at least thirty files, then ten iterative edits while tracking token consumption.
• Figma import of a real marketing page, including mobile breakpoints and asset weight under upload caps.
• Custom domain, SSL, and SEO metadata on Pro for a production-like hostname.
• Database persistence across redeploys, including backup and restore expectations.
• Auth flows: email login, OAuth if offered for your stack, and session expiry behavior.
• Web request volume under a simulated traffic spike relative to 333,000 Free or 1,000,000 Pro caps.
• Export or Git sync path to continue development outside Bolt.
• Teams seat with private NPM registry access for packages your org hosts internally.
• AI image editing output on Pro for hero and OG assets.

Analytics and operating visibility

Bolt.new is not a product analytics suite. Operating visibility for buyers means token usage dashboards, deploy logs, web request consumption against plan caps, and Teams billing views. During pilot, track tokens per editing session, tokens before and after GitHub import, and requests served once you share the preview URL widely.

Enterprise adds audit logs for organizations that must know who changed what. If compliance is the buying trigger, confirm log retention, export, and SIEM integration on the Enterprise order form, not from Free or Pro marketing copy. For growth teams, pair Bolt-hosted pages with your analytics stack early so marketing attribution does not depend on Bolt-native funnels that may not exist.

Security, data handling, and compliance

Bolt.new and StackBlitz publish security and commercial-use policies on official support pages. Enterprise buyers should expect SSO and audit logs as sales-led features. Free and Pro users still need to answer where customer data lives in Bolt-managed databases, who can access workspace projects on Teams, and whether generated code introduces vulnerable dependencies.

Reconcile marketing copy with your order form. Ask whether customer data in Bolt Cloud databases meets residency requirements, how deletion works when a project is removed, and whether NPM registry credentials on Teams are scoped per seat. If you import a GitHub repo, confirm you are not syncing secrets or private submodules you did not intend to expose to the builder environment.

Generated code can include API keys placeholders, open CORS rules, or permissive auth defaults. Treat every pilot deploy as untrusted until reviewed, the same way you would treat code from any AI assistant.

What questions should you ask before buying Bolt.new?

• What is the checkout price in our billing country, including annual versus monthly and taxes?
• How many tokens remain after importing a repo the size of our expected production app?
• Do Teams seats share tokens or meter independently per member?
• What happens when we exceed 1,000,000 web requests on Pro?
• Can we use our custom domain and SEO controls on Pro without Bolt branding?
• How do database backups, export, and deletion work on Bolt Cloud?
• What is the supported path to continue development in GitHub, Cursor, or Replit after export?
• Does Enterprise include SSO, audit logs, and a DPA suitable for our industry?
• Are private NPM packages on Teams scoped per user or org-wide?
• What upload size limit applies to our asset-heavy use case on Free versus Pro?

What red flags should you watch for with Bolt.new?

• Sales or demos only show greenfield apps with no GitHub import or large file tree.
• Teams is sold as a shared token pool when official pricing assigns tokens per member.
• Free hosting is presented as production-ready without discussing branding, request caps, or daily token limits.
• No one models web request limits before a marketing launch.
• Security is hand-waved because "it is just a prototype" while real customer data sits in Bolt databases.
• You needed deep IDE control on an existing monorepo and chose Bolt.new over Cursor or Replit without a portability test.

What are the best alternatives to Bolt.new?

Pick by where the product should live after day one, how you want to meter cost, and whether hosting is bundled or BYO. We only link tools we already cover on this site.

• LovableChoose Lovable when you want a similar prompt-to-hosted-app workflow with Lovable Cloud, Git sync, and credit economics that may fit your iteration style better than Bolt tokens.
• ReplitChoose Replit when a full cloud IDE, Replit Agent, deployments, and collaborative coding matter more than chat-only app generation inside Bolt.
• v0Choose v0 when the primary output is UI components and pages inside a Vercel-centric workflow rather than a fully hosted stack on Bolt Cloud.
• CursorChoose Cursor when you already have a repository, need local or hybrid IDE depth, and want agentic edits with usage pools tied to engineering seats instead of synced WebContainer trees.

Workflow test

What Bolt.new needs to prove in a real workflow

A clean landing page from a single prompt is not evidence. Run this four-step test on a project shape you actually plan to ship.

1. Match your file tree.Start one greenfield app and one GitHub import with similar features. Compare token burn over twenty edits on each.
2. Ship on Bolt Cloud.Deploy with auth and a database, attach a custom domain on Pro, and configure SEO for a real hostname.
3. Stress the caps.Model web requests against 333,000 Free or 1,000,000 Pro limits and upload assets near the 10MB or 100MB ceiling.
4. Exit and continue.Export or sync to GitHub, open the repo in Cursor or Replit, and confirm your team can maintain the app without re-scaffolding.

• Lovable reviewCompare prompt-to-app speed, Cloud hosting, and credit or token math side by side.
• Replit reviewCompare cloud IDE depth, Agent behavior, and deployment when Bolt feels too chat-centric.
• v0 reviewCompare UI generation and Vercel workflow fit when frontend output is the main deliverable.
• Cursor reviewCompare IDE-native agent editing when the repo already exists.
• AI app builders buyer guideCategory context for prompt builders, hosting bundles, and portability checks.
• How to choose an AI agent platformApply the same meter, data-use, and workflow criteria to app builders.
• Tools directoryScan adjacent builders, IDEs, and automation tools on one list.
• AI agent buyer scorecardTurn the four-step app builder test into a written go or no-go.

Claim and source ledger

Should you choose Bolt.new?

Bolt.new is a browser-based AI app builder with Bolt Cloud hosting for teams that want prompt-driven development, optional Figma and GitHub import, and published URLs without standing up infra first. Public pricing as of August 20, 2026 starts at $0 on Free with 1,000,000 monthly tokens and $25 per month on Pro with 10,000,000 monthly tokens. Teams is $30 per member per month with tokens that do not share across seats. Enterprise is custom for SSO and audit logs.

Choose Bolt.new when greenfield or modest-size apps, bundled hosting and databases, and chat-first iteration match your launch timeline, and you will model filesystem-driven token burn before you commit. Look elsewhere when Cursor or Replit fit an existing repo better, when v0 matches a Vercel UI workflow, when Lovable credit economics suit your team better, or when Enterprise SSO is mandatory and not yet quoted. Run the four-step workflow test on your file tree, not on a hello-world demo.