fused-overviewlisted

# What is Fused — and when should you use it Fused is a platform for running Python in the cloud, organized into projects and independently callable functions. The two core primitives are **canvases** and **UDFs**. --- ## The mental model **Canvas = project.** A canvas is the container for a piece of work — equivalent to a repo or a folder. All UDFs in a canvas share the same permissions and access controls (private / team / public). Set it once; all UDFs inherit it. **UDF = callable function with its own compute.** Each UDF (`@fused.udf def udf(...)`) is independently deployed, has its own API endpoint, its own compute, and its own cache. You can call one UDF without touching any other. Secrets and integrations (`fused.secrets`, `fused.api.notion_connect()`, etc.) are resolved automatically by the runtime — no credential management in code. ``` canvas (project) ├── udf_a.py → own endpoint, own compute, own cache ├── udf_b.py → own endpoint, own compute, own cache └── widget.json → optional browser UI on top ``` UDFs within a canvas call each other via `fused.load("udf_name")`, so you can compose them into pipelines while keeping each piece independently testable. --- ## When to use Fused instead of a local script | Situation | Use Fused? | |---|---| | Results need to be shared with others via a URL | ✓ | | Multiple people or systems need to call the same function | ✓ | | Work needs to run in parallel across many inputs | ✓ use `.map()` | | Compute