Prepare the Client and Starter Model
Use this setup before the TypeScript, Python, Go, .NET, or Swift quick start. The commands below assume a macOS or Linux shell. If you prefer an agent to handle setup, use Start from an Example.
1. Check your tools
The Cargo client calls TeaQL's remote evaluation and generation services for all seven languages. Using it to generate Python or TypeScript does not make your application a Rust application.
Install Rust and Cargo using the official Rust installer, then install the client version listed in Latest Versions:
cargo install cargo-teaql --version 2.0.14 --locked
cargo teaql --version
cargo teaql services
The first command should finish successfully; the second reports the installed client version. The third lists the targets available on the current service. Network access is needed for the client, generation service, and package registries.
Also install the toolchain for your chosen language:
| Language | Check before continuing | Generated dependency manifest |
|---|---|---|
| TypeScript | node --version and npm --version | package.json and its lock file |
| Python | python3 --version | pyproject.toml or the generated requirements file |
| Go | go version | go.mod |
| C# / .NET | dotnet --info | The generated project file and any global.json |
| Swift | swift --version | Package.swift |
Use the language version required by the generated manifest. A version used by an earlier example is not a universal minimum for every current target.
2. Download the school model
Create a new working directory and download the complete starter model:
mkdir teaql-school
cd teaql-school
mkdir models
curl --fail --location https://teaql.io/examples/school/model.xml --output models/main.xml
cargo teaql evaluate --input models/
The model describes a Platform, School Type constants (PRIMARY and
SECONDARY), and School. It comes from the generator's school-management test
fixture and includes a concrete package version. Keep models/main.xml as
the input for every generation and Assist command in the language guide.
Read the complete evaluation report. Resolve errors and repair or explicitly accept warnings before generation. A successful evaluation checks the model; it does not prove that a generated application builds or runs.
Your directory should now contain:
teaql-school/
models/
main.xml
Run subsequent generation commands from teaql-school/, so models/ and
generated/ refer to the same project root.
3. Choose a language and verify the result
Continue with TypeScript, Python, Go, .NET, or Swift. For Java and Rust, use their language guides.
The first successful run should produce a generated library and application, a passing build, one bounded query with comment and purpose, and one audited save followed by a readback. Keep the exact versions and commands with the results. If a step fails, start with Troubleshoot First Run.