Bolt.new Pricing vs Self-Hosted: What Indie Hackers Actually Pay

Bolt.new Pricing Tiers

Hidden Costs You Won't See on the Pricing Page

• ⚠ Bolt.new token consumption is opaque — complex multi-page builds drain tokens 3-10x faster than simple pages; you find out at runtime
• ⚠ Bolt.new does not include production hosting; deploy to Vercel ($0-20/mo), Netlify ($0-19/mo), or Cloudflare Pages ($0) separately
• ⚠ Bolt.new builds full-stack apps, not validation scaffolds — pSEO content collections, fake-door pricing, and post-signup survey must be added from scratch
• ⚠ Iteration is expensive — each round of changes to a Bolt-generated app costs tokens; validation involves a lot of iteration
• ⚠ Self-hosted requires CLI comfort for initial deploy; not a recurring cost, but a one-time time investment
• ⚠ Self-hosted with a framework like Astro requires knowing which packages to include — not zero-effort, but zero ongoing cost

The Token Model Explained

Bolt.new prices by token consumption, not by site or page. Every prompt, every code generation, every revision costs tokens. The free tier gives 150,000 tokens per day, which sounds generous until you’re iterating on a complex page layout across multiple sessions.

The Pro plan at $20/month gives 10 million tokens monthly. For a single validation site with moderate iteration, that’s usually sufficient. For an indie hacker running multiple experiments simultaneously, tokens run out faster than expected.

Self-Hosted Cloudflare: What It Actually Costs

A self-hosted Astro site on Cloudflare Pages has no hosting cost at validation-stage traffic. Cloudflare’s free tier covers:

• Unlimited static page deployments
• Cloudflare D1 (SQLite database) up to 5GB
• Workers API routes up to 100,000 requests per day
• Custom domain via your registrar (separate cost, typically $10-15/year)

The trade-off is setup time. Deploying an Astro site to Cloudflare requires running wrangler deploy and configuring a wrangler.toml. For most indie hackers who’ve deployed a side project before, this takes 30-60 minutes. It’s a one-time cost.

What’s Missing From Bolt.new for Validation

Three things that a validation site needs, which Bolt.new doesn’t include at any tier:

pSEO content collections. Organic traffic for validation comes from buyer-intent searches: “best alternative to [competitor]”, “[category] pricing”, “how to [task]”. These require content pages with proper frontmatter schemas, structured data, and internal linking. Bolt generates app routes, not content collections.

Fake-door pricing with click tracking. Bolt can generate a pricing page visually, but it won’t wire up click tracking to a database automatically. You need a component that records which tier was clicked, the page it came from, and a timestamp, then redirects to a waitlist. That’s custom code inside whatever Bolt generates.

Post-signup survey. Qualifying signups by role, current tool, and biggest pain point requires a survey step after email capture. Bolt doesn’t generate this workflow. You’d build it inside the generated app.

<DataTableBlock name=“Bolt.new vs Self-Hosted Cost Breakdown” description=“Monthly cost comparison for a validation site running on Bolt.new vs self-hosted Astro on Cloudflare” columns={[“Option”, “Monthly Cost”, “Hosting”, “pSEO”, “Fake-Door Pricing”, “Email Capture”]} rows={[ [“Bolt.new Free”, “$0 tokens + hosting”, “Separate”, “None”, “Build from scratch”, “Build from scratch”], [“Bolt.new Pro”, “$20 tokens + hosting”, “Separate”, “None”, “Build from scratch”, “Build from scratch”], [“Self-Hosted Cloudflare”, “$0”, “Included”, “Via Astro content collections”, “Via Validea scaffold”, “Via Validea scaffold”], [“Validea Starter”, “$9/mo”, “Included (Cloudflare)”, “Built-in”, “Built-in”, “Built-in”] ]} />

The Right Tool for Each Phase

Bolt.new and self-hosted Astro aren’t competitors for the same use case. They suit different phases:

Before you have demand signal: Use a self-hosted validation scaffold. Zero hosting cost, pSEO drives organic traffic, fake-door pricing tells you which tier to build around. Ship in a day.

After you have demand signal: Use Bolt.new to scaffold the actual product. You’ve validated the idea, you know which tier resonates, you know your buyer. Now Bolt’s full-stack generation saves real time building the real thing.

import AnswerBlock from ‘@validation/ui/seo/answer-block.astro’; import StatisticCitation from ‘@validation/ui/seo/statistic-citation.astro’; import DataTableBlock from ‘@validation/ui/seo/data-table-block.astro’;