Shakespeare
Shakespeare is an open-source AI app builder that runs entirely in your browser. Describe what you want, and it writes the code, previews it instantly and deploys it, while your project files stay on your own device.
What is Shakespeare?
Shakespeare is an open-source AI app builder that runs entirely inside a web browser. You describe the application or website you want in ordinary language, and the assistant writes the code, compiles it on the spot and shows a live preview. There is no account to create and nothing to install: opening the page is enough to begin.
What sets it apart is where the work actually happens. Shakespeare hosts no backend of its own beyond a few small, replaceable microservices. Project files live in the browser's own storage through LightningFS, Git runs client-side through isomorphic-git, and React and TypeScript are compiled locally with esbuild-wasm. The device talks straight to whichever AI provider the user has chosen, so the publisher never sees the code being written.
That choice of provider is deliberately left open. Thirteen providers ship preconfigured, among them OpenAI, Anthropic through OpenRouter, xAI, Z.ai, DeepSeek, Moonshot and Google, and a custom endpoint or a model running on the user's own machine can be added. The publisher also operates a default provider for people who would rather not manage an API key, billed from a credit balance topped up by card or over the Lightning Network, with several models free of charge.
The experience is not limited to chat. A file explorer and a syntax-highlighting editor allow any file to be changed by hand, assets can be dragged in, and a terminal is built in. Git is integrated rather than bolted on: commits are tracked automatically, any state can be rolled back in a single click, and projects push and pull to GitHub, GitLab or Nostr repositories. Finished work deploys to Netlify, Vercel, Cloudflare, Deno Deploy, Railway, nsite or the publisher's own hosting, and a full ZIP of the project can be downloaded at any moment for nothing.
Shakespeare is licensed under the AGPLv3 and published by Soapbox Technology, an organisation funded by grants and donations rather than by its users. It is specialised in, though not restricted to, applications built on the Nostr protocol.
What it does
- Turn a plain-language description into a working web application
- Preview the project instantly, compiled inside the browser
- Read and edit every file by hand in a built-in code editor
- Import an existing project from GitHub, GitLab, a Nostr repository or a ZIP archive
- Track changes with integrated Git and roll back to any earlier state in one click
- Deploy to Netlify, Vercel, Cloudflare, Deno Deploy, Railway, nsite or the publisher's hosting
- Download the complete source code as a ZIP archive at any time
When to use Shakespeare / When not to
A quick filter to help you decide if Shakespeare is the right fit.
When to use Shakespeare
- Non-technical people who want a working website or web app without writing a line of code
- Developers who want AI speed but refuse to give up direct access to the source and to Git
- Builders working on the Nostr protocol, which the tool is specifically tuned for
- Journalists, activists and anyone whose projects must stay private, since the code never leaves the device
- Makers on a tight budget who prefer paying per use, bringing their own API key or running a local model
When not to use Shakespeare
- Anyone who needs a backend service, a native mobile app or a desktop application, none of which the tool can build
- Teams handling sensitive or regulated personal data, which the publisher itself advises against
- Organisations that require a data processing agreement, a GDPR statement or a named legal entity before adopting a tool
- Users who want a managed cloud service with automatic backups and a support desk to call
- Buyers who need a fixed monthly subscription and a predictable invoice rather than usage-based spending
How to use Shakespeare
A typical end-to-end flow, from setup to results.
- Open shakespeare.diy in a browser; no installation and no account are required
- Read and accept the onboarding notice explaining that the tool runs on your device and is provided as is
- Choose an AI provider: your own API key, a Cashu token, an OAuth connection, or the publisher's default provider
- Pick a model from the list, where the cost per million tokens is shown next to each entry
- Create a new project, or import one from a Git URL, a nostr:// URI or a ZIP archive
- Describe what you want to build in the chat and let the assistant write and edit the files
- Use the Build button to compile and refresh the preview, and the Code and Logs views to inspect the result
- Edit any file by hand in the built-in editor and drag in your own images and assets
- Configure Git credentials if you want to push and pull, then commit and synchronise your work
- Deploy to a hosting provider, or export the whole project as a ZIP archive to keep a backup
Pros & Cons
Pros
- Genuinely open source under the AGPLv3, so the whole codebase can be inspected, modified and forked
- No vendor lock-in: the code sits on your device and a full ZIP export is always free
- Complete freedom of AI provider, including custom endpoints and models run locally
- No subscription; you pay only for the inference you actually consume, and free models exist
- Strong privacy by design, since the publisher states it cannot see the projects being built
- No account and no password required, with payment possible in Bitcoin for those outside the banking system
- Two-way Git support covering GitLab as well as GitHub, which few comparable tools offer
Cons
- No dedicated privacy policy, no GDPR mention, no data processing agreement and no subprocessor list
- No published legal entity or postal address; the non-profit structure was still being set up at the time of review
- Scope is limited to the browser: no backend services, no native mobile apps, no desktop applications
- Clearing browser data can destroy unsaved projects, and backups are explicitly the user's responsibility
- The service is provided as is, with no warranty and no guaranteed availability
- Credits are non-refundable once purchased, and model prices are set by third parties
- No consolidated pricing page and no dedicated support email; help runs through community channels
Pricing & Plans
Shakespeare provides a permanent free plan. The builder itself is open-source software and carries no paid tier, and it can be used indefinitely at no cost through the free models offered by the default provider, through a personal API key held with a third-party provider, or through a model run locally. No subscription of any kind is offered. The only chargeable element is AI inference through the publisher's default provider, billed as a pay-as-you-go credit balance topped up in United States dollars for an amount chosen by the user, with the processing fee displayed before payment. Published token rates begin at USD 1.50 per million input tokens for GLM-4.5. No monthly or annual entry price exists.
- Free — the builder is open-source software with no paid tier and no time limit
- Bring your own key — connect a personal account with any supported provider and pay that provider directly at its own rates
- Local models — run a model on your own machine
- such as Ollama
- GPT-OSS
- DeepSeek R1 or Gemma 3
- at no provider cost
- Shakespeare AI free models — several models consume no credits at all
- Shakespeare AI credits — pay-as-you-go balance topped up in USD by card or Lightning
- for a freely chosen amount
- with the processing fee shown before payment
- credits are non-refundable but never expire
- Shakespeare AI premium models — GLM-4.5 at USD 1.50 per million input tokens and USD 7.50 per million output tokens
- Claude Sonnet 4.5 at USD 4 and USD 16 respectively
- Gift cards — prepaid credit codes created in batches of one to one hundred
- shared as links or printable QR codes and redeemed into the recipient's balance
Data, GDPR & hosting
A consolidated view of how Shakespeare handles your data.
GDPR overview
There is no mention of the GDPR anywhere in Shakespeare's published material. A dedicated search across every collected page and both application bundles returned nothing: the regulation is never named, no compliance is claimed and none is denied. The tool has no standalone privacy policy either; the only privacy wording sits in section 10 of the Shakespeare AI terms, which states that authentication uses Nostr NIP-98 signatures, that no traditional account details or passwords are stored, that API requests may be logged for operational purposes, and that personal data is not sold or shared. No Article 27 EU representative is designated, no data processing agreement is offered and no subprocessor list is published. In practice the architecture collects very little, since project files stay in the user's browser, but that is a design consequence rather than a documented compliance position.
Who owns the data?
The terms assign ownership of AI-generated output to the user who requested it, to the greatest extent the law allows, and the user keeps every right over the prompts and messages they submit. The assignment carries reservations: the limits copyright law places on AI-generated works, the rights of third parties, and the terms of the underlying models. No warranty is given that the output is original or non-infringing. Project files themselves never reach the publisher, since they are held in the browser on the user's own device. The publisher reserves only the use of anonymised, aggregated data to improve its service, and expressly disclaims ownership of any specific input or output.
Reuse rights
Yes. Content generated in response to a request is assigned to the user, who may use, modify, distribute and commercialise it freely without seeking further permission. The complete source code of a project can be downloaded as a ZIP archive at any time and at no cost, deployed anywhere, developed further offline, or synchronised to a GitHub or GitLab account. The only limits are external rather than contractual: what copyright law recognises in AI-generated works, the rights of third parties, and the terms attached to whichever model produced the output. The publisher does not sell or share personal data with third parties.
Data retention & training
Hosting summary
Strictly speaking, Shakespeare hosts no user data at all. Project files, settings and API keys are held inside the user's own browser, in IndexedDB through LightningFS, and the publisher states that it stores neither personal data nor website content on its servers. That is why no hosting country and no hosting region are declared: there is nothing to locate. Persistent browser storage can be requested from the settings so that the browser does not purge the data when space runs low. What is hosted is the application itself. The shakespeare.diy domain resolves to a DigitalOcean address in the United States, and a handful of optional microservices run alongside it: a CORS proxy, a favicon service and the default AI provider. All three can be replaced in the settings. Two official mirrors exist for redundancy. If a user deploys a finished project, it goes to whichever host they selected, whether Netlify, Vercel, Cloudflare, Deno Deploy, Railway, nsite or the publisher's own hosting, and that host's jurisdiction then applies.
Things to keep in mind
Risks and trade-offs to weigh before adopting Shakespeare.
- Deleting your browser data can wipe out every project you have not exported or pushed to Git; the publisher puts backups squarely on you
- Handing a whole codebase to an assistant can erode your own grasp of it, leaving you unable to debug or audit what you have shipped
- The service is supplied as is, with no warranty and no guaranteed availability, and may be modified or discontinued at any time
- Once your prompts reach a third-party AI provider, that provider's terms govern them, not Shakespeare's
- The absence of a privacy policy, a GDPR statement or a data processing agreement is a real obstacle for any regulated activity
- Generated output carries no warranty of originality or non-infringement, so publishing it without review can create legal exposure
- Pay-as-you-go billing can drift upward through repeated debugging loops, and purchased credits are never refunded
Setup & Integrations
Technical difficulty
Very low. Opening the page is enough: nothing to install, no account and no password. The only mandatory steps are accepting the onboarding notice and choosing an AI provider and a model. The single point of friction is obtaining an API key from a third-party provider; users who prefer the built-in credits avoid it, though those require a Nostr login. Git credentials are optional and only needed for synchronisation. Advanced settings exist for the CDN, CORS proxy, relays and system prompt, but none of them has to be touched.
Deployment
Integrations
Supported languages
Behind Shakespeare
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Shakespeare.
Frequently asked questions
What is Shakespeare and how does it work?
Do I need to know how to code?
How much does it cost?
What can I build with it, and what can I not build?
Which AI providers can I use?
Where is my data stored?
Can I get my source code out?
Can I import a project I have already started?
What should I be careful about?
Should you pick Shakespeare?
Shakespeare occupies a position almost no other AI builder does: it is genuinely open source under the AGPLv3 and it runs without a backend, which turns its privacy claims into something structural rather than promissory. Because project files never leave the browser, the publisher's assertion that it cannot see what you are building is a consequence of the architecture, not a policy that could quietly change. The same design gives the tool its most distinctive quality: the complete absence of lock-in. Any project can be exported as a ZIP archive at any moment, pushed to GitHub or GitLab, and continued elsewhere.
The trade-off is equally clear. Nothing is managed for you. There is no cloud safety net, backups are your responsibility, and clearing your browser data can destroy unsaved work. The legal paperwork is unusually thin for a published tool: no privacy policy, no GDPR statement, no data processing agreement, no postal address and no incorporated entity, since the non-profit structure was still being established at the time of review. Organisations with procurement or compliance requirements will find that hard to accept, however sound the engineering is.
The economics are unusual too. The publisher takes no revenue from its products or its users, living on grants and donations, and the only money that changes hands pays for AI inference on a pay-as-you-go basis. That removes the pressure to monetise attention or data, but it also means the tool's future rests on continued philanthropic funding.
For hobbyists, freelancers, Nostr builders, and anyone who values sovereignty over convenience, Shakespeare is a compelling and honest piece of software. For regulated businesses, it is not yet a realistic option.
- Choosing a selection results in a full page refresh.
- Opens in a new window.