Development workflow¶

A Cobra4 project usually starts with a template, then uses the same compiler path from the CLI, Studio, tests, and CI.

Start a project¶

c4 init my-service --template http-service
cd my-service
c4 doctor
c4 studio

Useful templates:

Template	Use it for
`http-service`	API handlers and `serve` workflows.
`etl-pipeline`	`read` / transform / `save` data jobs.
`daemon`	`every`, `on event from`, scheduler, and queue work.
`agent`	LLM agent examples with a mock provider.

List templates with:

c4 init --list

Daily loop¶

c4 check src/main.c4 --strict
c4 fmt src/main.c4 -w
c4 run src/main.c4
c4 test

Use c4 run FILE --watch while iterating on one entry point. Use c4 studio when you want the project tree, snippets, terminal, graph, and generated Python/source-map view in one place.

Inspect generated Python¶

c4 build src/main.c4 -o build/main.py --source-map

Studio shows the same generated Python interactively and highlights the mapping between Cobra4 source lines and generated Python lines.

Dependencies and plugins¶

cobra4.toml is the project root signal. It can declare runtime dependencies, language plugins, secrets backend, and host inventory.

c4 deps add requests --version 2.31.0
c4 deps install --venv
c4 plugin list
c4 doctor

c4 doctor checks declared dependencies and configured language plugins, then reports missing optional extras as informational warnings.

Daemons and services¶

For long-running files that contain every, on event from, or serve handlers:

c4 serve src/main.c4

During development you can still smoke-test the same file with:

c4 run src/main.c4
c4 check src/main.c4 --strict

CI baseline¶

A minimal CI job should run:

c4 doctor
c4 check src/main.c4 --strict
c4 test

For libraries or deployment scripts, also build the generated Python:

c4 build src/main.c4 -o build/main.py --source-map