docs/API_REFERENCE.md

# API Reference

This document describes the public Python modules and functions available in AnyCoder.

---

## `models.py`

### `ModelInfo` dataclass

```python
@dataclass
class ModelInfo:
    name: str
    id: str
    description: str
    default_provider: str = "auto"
```

### `AVAILABLE_MODELS: List[ModelInfo]`

A list of supported models with metadata.

### `find_model(identifier: str) -> Optional[ModelInfo]`

Lookup a model by name or ID. Returns a `ModelInfo` or `None`.

---

## `inference.py`

### `chat_completion(model_id: str, messages: List[Dict[str,str]], provider: Optional[str]=None, max_tokens: int=4096) -> str`

Send a one-shot chat completion request. Returns the assistant response as a string.

### `stream_chat_completion(model_id: str, messages: List[Dict[str,str]], provider: Optional[str]=None, max_tokens: int=4096) -> Generator[str]`

Stream partial generation results, yielding content chunks.

---

## `hf_client.py`

### `get_inference_client(model_id: str, provider: str="auto") -> InferenceClient`

Create and return a configured `InferenceClient`, routing to Groq, OpenAI, Gemini, Fireworks, or HF as needed.

---

## `deploy.py`

### `send_to_sandbox(code: str) -> str`

Wrap HTML code in a sandboxed iframe via a data URI for live preview.

### `load_project_from_url(url: str) -> Tuple[str, str]`

Import a Hugging Face Space by URL, returning status message and code content.

---

## `plugins.py`

### `PluginManager`

* `discover()`: auto-discovers plugins in the `plugins/` namespace.
* `list_plugins() -> List[str]`: return registered plugin names.
* `run_plugin(name: str, payload: Dict) -> Any`: execute a plugin action.