bolt.new — AI Readiness Score: 60/100

doclight — ai Readiness Report

bolt.new

Pages analyzed (12)

llms.txt included https://bolt.new/llms.txt 200
llms-full.txt included https://bolt.new/llms-full.txt 200
marketing https://bolt.new/ 200
pricing included https://bolt.new/pricing 200
pricing included https://support.bolt.new/account-and-subscription/team-plans 200
support included https://support.bolt.new/ 200
support included https://support.bolt.new/integrations/git 200
support included https://support.bolt.new/integrations/expo 200
support included https://support.bolt.new/integrations/figma 200
marketing https://status.bolt.new/ 200
other https://bolt.new/figma 200
other https://status.bolt.new/history 200

llms.txt found

AI READINESS SCORE

Bolt (bolt.new) is an AI-powered website and app builder by StackBlitz that lets users prompt, run, edit, and publish web and mobile apps directly in the browser. Users describe what they want to build in natural language and Bolt generates the code, with integrations for Figma, Expo, GitHub, Supabase, Stripe, Netlify, and MCP servers. It includes hosting, databases, custom domains, and team collaboration features, with usage metered in AI tokens.

Dimension Breakdown

Discoverability

The product is clearly named, described, and indexed with an llms.txt documentation index, making it easy to find and categorize.

Comprehension

It is very clear that Bolt is an AI app/website builder, with extensive prose explaining its capabilities and integrations.

Setup clarity

Getting started is conversational (prompt-based) and documented for integrations, but no programmatic onboarding or account setup steps are exposed for an agent.

Documentation

Rich human-readable documentation exists with quickstarts, tutorials, and integration guides, plus a markdown-friendly llms.txt index.

Pricing clarity

Pricing tiers, token limits, and feature differences are clearly laid out, though token-to-cost conversion remains somewhat ambiguous.

Integration examples

Tool integrations (Figma, Expo, GitHub, Stripe) are documented as UI workflows, but there is no public API, SDK, or code examples for programmatic use.

Agent Journey

✓

Discover

✓

Understand

Setup

Setup is entirely UI/chat-driven with no programmatic account creation or API key flow, so an autonomous agent cannot self-onboard without a human.

Use

There is no public API or OpenAPI spec, so an agent cannot programmatically invoke Bolt beyond a human-facing chat interface.

Confusion Points

med The llms.txt and llms-full.txt URLs returned HTML page shells rather than the expected plain-text documentation index https://bolt.new/llms.txt and /llms-full.txt
med Token-based pricing is hard to map to actual costs or app sizes, making programmatic cost estimation difficult pricing page FAQ on tokens
high No API, SDK, or authentication documentation for programmatic interaction; everything is described as in-app UI actions support/docs across integration pages

Missing Information

Is there a public REST/GraphQL API or SDK to interact with Bolt programmatically?
How can an agent authenticate (API keys, OAuth) without manual browser login?
What is the exact token-to-dollar or token-to-action conversion for budgeting?
Is there an OpenAPI/Swagger specification available?
Can projects be created, built, and published entirely via API rather than the chat UI?
What are the rate limits and quotas for programmatic access?

Improvements

Serve llms.txt and llms-full.txt as actual plain-text/markdown content instead of an HTML app shell so agents can parse them.
Publish a public API reference with an OpenAPI/Swagger spec covering project creation, build, and publish operations.
Document an authentication flow (API keys/OAuth) that allows non-interactive, agent-driven access.
Provide code examples and SDK snippets for common operations and integrations (GitHub, Supabase, Stripe).
Add a clear token-cost calculator or worked examples mapping token usage to project size and dollar cost.
Include machine-readable pricing data (e.g., JSON) so agents can parse and compare plans deterministically.