ShoalStack is now
ShoalStack is now
Tell the built-in platform201 AI assistant what your backend needs, and it adds each API, worker, database, and queue to a visual graph of your cloud infrastructure.
platform201 shows your backend as a visual graph of services and their connections. When you name a service, the AI assistant adds it to the graph, connects it, and gives it the settings it needs.
Ask for a Python API, a Postgres database, a queue, or another supported service. The assistant adds each one to the graph as a separate node.
Connect services to databases, buckets, queues, and gateways. Each connection gives a value, such as a connection string, to a variable in your code.
Give each node its language, port, and region. You can also set a gateway route, the cron expression of a scheduler, or the version of a database.
Upload source code or attach a repository to a container or function node. Then ask the assistant for the route and database of that service.
Generation is most useful when you know the shape of the system before you build it.
The code may not exist yet. A known pattern may fit. A working API may need one more service.
Plan a frontend, an API, a queue, and a database. Check their routes before you write code.
Start from Full Stack Next.js Application.
Ask the assistant to change its nodes and connections.
Add a queue and a worker container. Connect the worker to the database your API uses.
This example builds a research agent that answers questions, stores uploads in a private bucket,
and saves conversations. Each step is one request to the platform201 AI assistant.
Ask the platform201 AI assistant: “Build a Python API on port 8080 behind the shared gateway at /api.” The graph shows a gateway and a container.
Ask: “Add a private bucket, a Cloud SQL database, and a daily job that reindexes the documents.” The assistant connects all three to the API.
Ask: “Attach my Python source and add my AI and search API keys, but do not deploy yet.” The container now holds your code and keys.
Ask: “Check the graph, then deploy it.” When the deployment status reads succeeded, the agent answers at the public URL of its gateway.
A generated graph is the same as a graph you build by hand. You can edit it in the graph editor, run its services as one application, or give the assistant more requests as the project grows.
Ask for one Python container behind the shared gateway. Then add a Cloud SQL database, a queue, or a worker. Deploy when the graph and the source code are ready.
Yes. An AI assistant can turn a description of the services you need into infrastructure configuration, such as nodes, connections, and settings. The result still needs review, because an assistant can choose a wrong port, route, or setting. The platform201 AI assistant adds each service to a visual graph as a node, connects it, and gives it settings.
You can name all the services in one request, or add them one at a time. Either way, the assistant builds the graph through a series of separate changes, so check the result before you deploy.
