10 apps you can build in Studio, no code

People ask us what Digitorn is for. The honest answer is "applications". Not just chatbots, not just coding agents, not just RAG. The workspace is generic. Anything that can be expressed as a loop of LLM decisions plus tool calls fits the model, and you build it by describing it in Studio, not by writing code.

To make that concrete, here are ten very different apps, each one a real shape we've built or seen built on the platform. You assemble any of them in Studio with the builder assistant, or install a ready-made version from the Hub.

At the centre is your agent. The orbits are categories of work the runtime handles natively (developer tools, live apps, knowledge work, messaging, background jobs). Pick any one, describe it, ship it.

#A coding agent that behaves like Claude Code

The flagship use case. A coordinator on a strong model that plans, edits files, runs tests, and dispatches search work to a cheap explorer worker. We documented the full architecture in the Claude Code clone post, but the idea is small: one planner, one fast worker, a coding safety profile, and a per-agent model split that keeps the bill down. Install the ready-made digitorn-code from the Hub, or build your own in Studio (cost analysis).

#A website builder that edits in front of you

This is Sites, Digitorn's website builder, our answer to Lovable. You describe the site you want, the agent writes the code into your workspace, and a live preview shows the result next to the chat as edits land. You keep talking to it ("make the hero darker", "add a pricing section") and watch the page change in real time. When it's ready, you publish it. No code, no framework, no setup.

#A research agent that cites its sources

Half the value of a research bot is not making things up. The shape that holds up: one coordinator that plans and writes, one or two researchers that fetch sources, and a fact-checker on zero temperature that verifies claims before the writer ships the answer. Most of the token spend lands on a cheap model, with citations the writer can't bypass. The ready-to-use version is the deep research template.

#A LaTeX writer that compiles as you type

Point an agent at a .tex document and give it a PDF preview pane. It edits the source through the chat, the runtime invokes the LaTeX toolchain, and the PDF re-renders as writes land. Useful for paper drafts, technical docs, and anything that benefits from real typography. This is the shape behind Digitorn's own LaTeX builtin, and you get inline error feedback on every save, the same way you'd get TypeScript errors in a code editor.

#A Discord bot that answers in a channel

Connect Discord and any agent becomes a bot: pick the adapter, connect the account, and name the agent that wakes up when a message arrives. The bot token is never exposed; it's resolved from the encrypted vault at run time. The same shape works for Telegram, Slack, and the rest, you just pick a different channel.

#A scheduled report that runs every Monday morning

Triggers turn an agent into a worker that fires on a schedule, on a webhook, or on an event. It runs to completion, posts the result somewhere (Slack, email, storage), and exits. No always-on process to babysit. We use this internally for a "what shipped last week" digest: the agent reads our changelog, summarises, and posts it onward. You set the schedule right where you build the agent.

#A pull-request triager that catches itself drifting

Webhooks are a schedule's cousin. A service like GitHub fires the agent on each new PR, and it runs against the diff. Add a guardrail that caps how many tool calls a single run can make, and you have a triager that can't accidentally cost you a fortune on an unusual PR. Past the ceiling, the run stops and you decide whether to extend.

#A documentation generator that updates with the code

Point it at a directory and let it read the source plus the existing docs. It writes real Markdown, and a live preview shows each page as it's written. The coding safety profile's read-before-edit rule means you can trust that it read a module before describing it. It's a more forgiving cousin of the coding agent, same idea, narrower scope. The ready-to-use version is the doc writer template.

#A slide builder that ships an interactive deck

This is Slides, Digitorn's presentation agent. You describe the deck ("a 10-slide pitch for a B2B analytics tool") and it builds a real, interactive presentation you refine by chatting at it, one idea per slide. Useful for kickoff decks, recap meetings, and pitch drafts. It's surprising how much faster it is to shape a deck by talking to it than by clicking through Keynote.

#A knowledge agent that answers from your own documents

Point this one at a folder of documents instead of a live web search, and it answers the way a support engineer who actually read them would, citing the source instead of guessing. Embedding, chunking, and citations all have sane defaults, so it works out of the box, and you can point it at a real folder or let the agent load documents itself. The ready-to-use version is the knowledge base agent template.

#What ties them together

Ten apps. Different tools, different interfaces, different cadences. The shape is the same in every one: choose what the agent has access to, choose its brain, choose how it's triggered, and tell it what to do. Everything else (orchestration, abort handling, live reload, marketplace packaging, credentials, audit) is the runtime's job.

That's the bet behind the platform. Most of what people end up writing in a code framework is plumbing the framework should have handled. Push the plumbing into the runtime, leave the person with a workspace, and the time-to-app collapses.

If one of these shapes lines up with something you're building, the fastest way to start is the templates page: ten patterns, each one ready to open in Studio and adapt.

1. Open Studio in your browser, nothing to install.
2. Describe the app you want, or open a template.
3. The builder assistant assembles it and you refine it live.

Digitorn is open source on GitHub. If you build something interesting, publish it to the Hub. That's how the catalogue grows.

#Further reading

If you want to dig into specific patterns:

• The cost trick that makes multi-agent setups affordable: How we cut our coding agent's bill by 60%
• Why you shouldn't write your agents in code: Why you shouldn't write your agents in code
• The full architecture of the coding agent: How to build your own Claude Code
• Foundations if you're new: What is an AI agent