feat(setup): provider registry and interactive setup wizard

- setup/providers.py: registry for 8 providers (3 local, 5 cloud), frozen dataclasses
- setup/wizard.py: questionary-based wizard — provider select, model input,
  API key collected via vault.writer (not config.yaml), connectivity check,
  test call via litellm

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

src/pyra/setup/providers.py

@@ -0,0 +1,104 @@
+from dataclasses import dataclass, field
+
+
+@dataclass(frozen=True)
+class Provider:
+    id: str
+    display_name: str
+    requires_key: bool
+    default_model: str
+    litellm_prefix: str
+    base_url: str | None = None
+    key_env_var: str | None = None
+    connectivity_check: str | None = None
+    group: str = "Cloud"
+
+
+PROVIDERS: list[Provider] = [
+    # ── Local ────────────────────────────────────────────────────────────────
+    Provider(
+        id="lmstudio",
+        display_name="LM Studio (local)",
+        requires_key=False,
+        default_model="local-model",
+        litellm_prefix="openai/",
+        base_url="http://localhost:1234/v1",
+        connectivity_check="http://localhost:1234/v1/models",
+        group="Local",
+    ),
+    Provider(
+        id="ollama",
+        display_name="Ollama (local)",
+        requires_key=False,
+        default_model="llama3",
+        litellm_prefix="ollama/",
+        base_url="http://localhost:11434",
+        connectivity_check="http://localhost:11434/api/tags",
+        group="Local",
+    ),
+    Provider(
+        id="llamacpp",
+        display_name="llama.cpp server (local)",
+        requires_key=False,
+        default_model="local-model",
+        litellm_prefix="openai/",
+        base_url="http://localhost:8080/v1",
+        connectivity_check="http://localhost:8080/v1/models",
+        group="Local",
+    ),
+    # ── Cloud ────────────────────────────────────────────────────────────────
+    Provider(
+        id="anthropic",
+        display_name="Anthropic (Claude)",
+        requires_key=True,
+        default_model="claude-sonnet-4-6",
+        litellm_prefix="anthropic/",
+        key_env_var="ANTHROPIC_API_KEY",
+        group="Cloud",
+    ),
+    Provider(
+        id="openai",
+        display_name="OpenAI (GPT)",
+        requires_key=True,
+        default_model="gpt-4o",
+        litellm_prefix="openai/",
+        key_env_var="OPENAI_API_KEY",
+        group="Cloud",
+    ),
+    Provider(
+        id="gemini",
+        display_name="Google (Gemini)",
+        requires_key=True,
+        default_model="gemini/gemini-2.0-flash",
+        litellm_prefix="gemini/",
+        key_env_var="GEMINI_API_KEY",
+        group="Cloud",
+    ),
+    Provider(
+        id="deepseek",
+        display_name="DeepSeek",
+        requires_key=True,
+        default_model="deepseek/deepseek-chat",
+        litellm_prefix="deepseek/",
+        key_env_var="DEEPSEEK_API_KEY",
+        group="Cloud",
+    ),
+    Provider(
+        id="qwen",
+        display_name="Qwen (Alibaba)",
+        requires_key=True,
+        default_model="openai/qwen-plus",
+        litellm_prefix="openai/",
+        base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",
+        key_env_var="DASHSCOPE_API_KEY",
+        group="Cloud",
+    ),
+]
+
+PROVIDERS_BY_ID: dict[str, Provider] = {p.id: p for p in PROVIDERS}
+
+
+def get_provider(provider_id: str) -> Provider:
+    if provider_id not in PROVIDERS_BY_ID:
+        raise KeyError(f"Unknown provider: {provider_id!r}")
+    return PROVIDERS_BY_ID[provider_id]

src/pyra/setup/wizard.py

@@ -0,0 +1,139 @@
+import httpx
+import questionary
+from rich.console import Console
+from rich.panel import Panel
+from rich.text import Text
+
+from pyra.config.manager import save_config
+from pyra.config.schema import ProviderConfig, PyraConfig
+from pyra.setup.providers import PROVIDERS, Provider, get_provider
+
+console = Console()
+
+
+def run_setup() -> None:
+    console.print(Panel(
+        Text("Welcome to Pyra Setup", justify="center", style="bold cyan"),
+        subtitle="Personal AI Assistant",
+        border_style="cyan",
+    ))
+    console.print()
+
+    provider = _choose_provider()
+    model = _choose_model(provider)
+
+    if provider.requires_key:
+        _collect_api_key(provider)
+
+    _test_connection(provider, model)
+
+    cfg = PyraConfig(
+        ai=ProviderConfig(
+            provider_id=provider.id,
+            model=model,
+            base_url=provider.base_url,
+        )
+    )
+    save_config(cfg)
+
+    console.print()
+    console.print(Panel(
+        f"[green]Setup complete![/green]\n\n"
+        f"Provider: [bold]{provider.display_name}[/bold]\n"
+        f"Model:    [bold]{model}[/bold]\n\n"
+        "Run [bold cyan]pyra chat[/bold cyan] to start talking.",
+        border_style="green",
+    ))
+
+
+def _choose_provider() -> Provider:
+    local = [p for p in PROVIDERS if p.group == "Local"]
+    cloud = [p for p in PROVIDERS if p.group == "Cloud"]
+
+    choices = (
+        [questionary.Choice("── Local ──────────────────", disabled=True)]
+        + [questionary.Choice(p.display_name, value=p.id) for p in local]
+        + [questionary.Choice("── Cloud ──────────────────", disabled=True)]
+        + [questionary.Choice(p.display_name, value=p.id) for p in cloud]
+    )
+
+    provider_id = questionary.select(
+        "Choose your AI provider:",
+        choices=choices,
+    ).ask()
+
+    if provider_id is None:
+        raise SystemExit(0)
+
+    provider = get_provider(provider_id)
+
+    if provider.connectivity_check:
+        _check_local_server(provider)
+
+    return provider
+
+
+def _check_local_server(provider: Provider) -> None:
+    console.print(f"  Checking connection to [bold]{provider.display_name}[/bold]...", end=" ")
+    try:
+        resp = httpx.get(provider.connectivity_check, timeout=3.0)
+        resp.raise_for_status()
+        console.print("[green]✓[/green]")
+    except Exception:
+        console.print("[yellow]✗ (server not reachable)[/yellow]")
+        console.print(
+            f"  [yellow]Warning:[/yellow] Could not reach {provider.base_url}.\n"
+            f"  Make sure {provider.display_name} is running before using Pyra."
+        )
+
+
+def _choose_model(provider: Provider) -> str:
+    model = questionary.text(
+        "Model name:",
+        default=provider.default_model,
+    ).ask()
+    if model is None:
+        raise SystemExit(0)
+    return model.strip()
+
+
+def _collect_api_key(provider: Provider) -> None:
+    from pyra.vault.writer import set_key
+
+    console.print(
+        f"\n  [dim]API key will be stored in the encrypted vault — never in config.yaml[/dim]"
+    )
+    key = questionary.password(f"Enter your {provider.display_name} API key:").ask()
+    if key is None:
+        raise SystemExit(0)
+    key = key.strip()
+    if not key:
+        console.print("[red]No key entered — skipping.[/red]")
+        return
+    set_key(provider.id, key)
+    console.print("  [green]✓ Key stored in vault[/green]")
+
+
+def _test_connection(provider: Provider, model: str) -> None:
+    from pyra.vault.reader import get_key
+
+    console.print("\n  Running test call...", end=" ")
+    try:
+        import litellm
+
+        api_key = get_key(provider.id) if provider.requires_key else "no-key"
+        kwargs: dict = {
+            "model": f"{provider.litellm_prefix}{model}",
+            "messages": [{"role": "user", "content": "Reply with exactly: OK"}],
+            "max_tokens": 10,
+        }
+        if provider.base_url:
+            kwargs["api_base"] = provider.base_url
+        if api_key and api_key != "no-key":
+            kwargs["api_key"] = api_key
+
+        litellm.completion(**kwargs)
+        console.print("[green]✓ Connection OK[/green]")
+    except Exception as exc:
+        console.print(f"[yellow]✗ Test call failed: {exc}[/yellow]")
+        console.print("  [dim]You can still proceed — check your config with 'pyra setup' again.[/dim]")