Turn this week’s research into a usable workspace
✦
Sigla Agent
The first version is ready — sources, views, and automations are all in it. Start using it.
Building the research workspaceReady to use
✓Add a daily refresh…↗
SIGLA · AI AGENT SPACE
No code? Let AI build the app you need.
Sigla is a space where AI builds the software you need. Describe the app. AI builds it, connects Agents, Skills, and MCP, and puts it to work. Research, knowledge, email, and automation can all start the same way.
WHAT IS SIGLA
Sigla is an AI Agent Space.
An AI Agent Space is an AI-native workspace where you describe the app or workflow you need, AI builds it with MCL, and both you and your agents use it. The code doubles as context, so AI can read, change, run, and verify the result. No code, no deployment. An app for agents, not code you maintain.
02 · START WITH YOUR PROBLEM
No code, missing features, painful integrations, or software that’s a chore to use? Space handles it.
Sigla doesn’t ask you to learn another tool. It takes over the hard parts — building, deploying, connecting capabilities, changing features, and the daily friction of using the software. You describe what you need; AI handles the rest.
THE PROBLEMS
THE SOLUTIONS
AI-NATIVE SPACE
Sigla Space
You want an app — but can’t code or deploy
Tell AI what you want to build. It builds the app, and it’s ready to use — no programming and no deployment.
HOW IT DIFFERS
Chatbot, low-code, coding agent — or an AI Agent Space?
A category-level comparison. Individual products vary.
| Dimension | Chatbot | Low-code / automation | Coding agent / IDE | Sigla AI Agent Space |
|---|---|---|---|---|
| Where the result lives | In the conversation | In a flow you assembled | In a code repository | In a runnable Space |
| Who can use it | People typing | People running the flow | Developers | People and agents |
| Can AI change it in place? | No | Limited | Yes, in code | Yes, in the same Space |
| Do you write code? | No | Configuration only | Yes | No |
| Can AI operate it for you? | No | Partly | No | Yes |
| What you start with | A prompt | A blank canvas | A repository | A request or a template |
03 · WHY IT WORKS
Code is context. Context is code.
These capabilities don’t come from stacking features. They come from MCL — the Metaverse Construction Language — an AI-native programming framework for AI programming and AI applications. The code is the context: interface, data, logic, and capability access are all expressed in code that explains itself. AI can read, change, run, and verify the app directly. When it’s built, it’s immediately usable — by you and by your Agent.
Learn about MCL →MCL code (and context)
<view>
[top_session]
[h1]
{
"id": "page_title",
"text": "todo"
}
[button]
{
"id": "refresh_btn",
"text": "refresh",
"onClick": "handle_refresh"
}
{when: self.is_loading}
[loading_indicator]
{
"text": "loading..."
}
{when: not self.is_loading}
[task_list]
{each: self.todo_list as task}
[task_item]
[text]
{
"text": "$task.task_name"
}
</view>
<style>
[top_session]
display: flex
justify-content: space-between
align-items: center
margin-bottom: 16px
{"id":"page_title"}
font-size: 24px
font-weight: bold
color: #333
{"id":"refresh_btn"}
background-color: #007bff
color: white
border-radius: 4px
padding: 8px 16px
[loading_indicator]
color: #666
[task_list]
display: grid
gap: 12px
[task_item]
padding: 12px
border-radius: 6px
background-color: #f7f7f7
</style>
<script>
state:
todo_list: []
is_loading: False
def init:
@flow:
self.fetch_todo_list()
def fetch_todo_list:
@desc: "Fetch the todo list"
@flow:
self.is_loading = True
self.todo_list = await mcs.fetch_todo_list(page=1, limit=10)
self.is_loading = False
def handle_refresh:
@desc: "Refresh the todo list"
@flow:
self.fetch_todo_list()
</script>The same code, running
Code that explains itself
MCL expresses interface, data, logic, files, Skills, and automations as one readable body of code. An Agent can see what the app does and where its state lives — no extra docs, no reverse engineering. Agent, Skill, and MCP access live in the same code.
Built for AI to program
MCL is designed for AI to read, change, run, and verify. AI can understand the app at lower cost, and complex changes become something it can handle — lowering the difficulty of building apps.
Built to be used — by you and your Agent
MCL code maps directly to the running app. It’s usable as soon as it’s built, with no separate deployment. You use the interface; your Agent can read state, call capabilities, and act for you.
04 · WHAT YOU GET
Ask for what you need. AI builds it — and uses it for you.
What you get is more than an app. It’s a new way of working: a lower barrier to building, capabilities that assemble around your needs, and software that’s usable the moment it’s built. AI doesn’t stop at building — it uses the software for you. In Sigla, building and using aren’t two phases; they’re two ends of the same conversation.
If you can say it, you can build it
No programming or deployment. Describe what you need in everyday language; AI builds the first version and runs it in front of you.
Connect and assemble what you need
Agents, Skills, MCP, and external plugins connect to the same Space and combine around your work — no separate system for every capability.
Features grow with your needs
When something is missing, just ask. AI changes the code and adds the feature to the existing app — no rebuild, no waiting for a release.
AI builds it — and uses it for you
From the first version and later feature changes to the daily work inside the app, AI can keep going. You don’t have to switch between being the developer and being the operator.
05 · HOW TO START
Start with an idea or a template — and get a working first version
Start with a sentence about what you need, or with a mature template. Once the first version is ready, use it in the same Space — and ask AI to change whatever is missing.
Start with an idea
Tell AI what you want to build and what work it should do. It builds a usable first version.
Create an AI Agent SpaceStart with a template
Load a research, knowledge, email, content, investment, or automation template, then adapt it to your work.
Browse workspace templates01
Email Management
The email management center based on the email plug-in allows AI to help you manage contacts efficiently and extract information from important emails, such as bill management and important matter management.
02Investment Analysis
AI investment research workbench: multi-market monitoring, credibility-scored news, and four-layer framework reports centered on expectation gaps — with automatic framework evolution after every analysis.
03Knowledge Base
Personal knowledge base management tool - let AI manage documents and forms. Is the data scattered on the network disk, chat history, or desktop? Import and parse md/pdf/doc/docx/xlsx/xls, and Excel will directly turn it into an editable table; no need to copy and paste every cell. Reimbursement ledgers, project materials, and study notes can all be filed like this - you just say what you want and leave the rest to it.
04Research Space
A workspace for collecting, categorizing, and analyzing research content.
Common questions about building and using apps with AI
Can I build an app without knowing how to code?
Yes. Describe what you need in everyday language. AI builds it with MCL, and it’s ready to use — no programming or deployment required.
Do I need to deploy the app myself?
No. MCL code maps directly to the running app. When development is done, it’s usable in the Space.
What if the app is missing a feature I need?
Ask AI. It changes the code and adds the feature to the existing app — no switching tools and no waiting for a release.
How do Agents, Skills, and MCP connect to a Space?
MCL has built-in access for all three. One line of code connects them, and AI can handle the integration for you.
Can AI use the software it builds for me?
Yes. The code is the context. Your Agent reads the same interface, data, and state, calls the app’s capabilities, and operates the app for you.
How is Sigla different from ChatGPT or a low-code platform?
ChatGPT-style tools mainly return conversation. Low-code platforms still ask you to assemble flows and components. Sigla has AI write the code directly, so the app can keep changing, the result stays in one Space, and both you and your Agent can use it.
Will AI break the app when it changes it?
Changes happen in the same Space. AI reads the existing code before changing it, and you can check the result in the interface right away. If something is off, ask AI to adjust it.
How long does it take to learn?
If you can describe what you need, you can start. AI builds the first version, so there’s no development tool to learn first — just keep telling AI what to change.
What is an AI Agent Space?
It is a workspace where people and AI agents view, use, change, and keep working on the same app or workflow. It holds an interface, data, files, logic, and connected capabilities. The conversation is where you ask; the result stays in the Space.
What is AI Native, and what is an AI Native Framework?
AI Native means designing software around what AI can read, change, run, and verify. An AI Native Framework such as MCL makes the code the context, so an agent can understand and change the app directly.
Is Sigla a vibe coding tool?
It shares the conversational starting point of vibe coding: describe what you want and iterate. The difference is that the result stays in a runnable Space, the code is readable context, and both people and agents can keep changing and using it.
What is an App for Agent?
An App for Agent is software designed for an AI agent as well as a person. Its interface, data, and capabilities are exposed as readable context, so an agent can operate it without a human clicking through every step. Sigla Spaces are built this way.
What is MCL?
MCL (Metaverse Construction Language) is Sigla's AI Native Framework and Space language. It expresses interface, data, logic, files, Skills, and automations as one readable body of code. It is already used in Sigla; the open-source release is in preparation.
Can I start from a template and change it later?
Yes. A template is a working first version, not a fixed product. After loading it, you can ask AI to change the interface, data, or automations, and save the result as your own template.