feat: add multi-image vision analysis

apps/ai/providers/base.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 from dataclasses import dataclass, field
-from typing import Any, Mapping, Protocol
+from typing import Any, Mapping, Protocol, Sequence
 
 
 class AiProviderError(RuntimeError):
@@ -78,6 +78,12 @@ class ImageGenerationResult:
     raw: Mapping[str, Any]
 
 
+@dataclass(frozen=True)
+class MultimodalImage:
+    data: bytes
+    mime_type: str = "image/png"
+
+
 class Provider(Protocol):
     def capabilities(self) -> set[str]:
         ...
@@ -108,6 +114,16 @@ class Provider(Protocol):
     ) -> ImageGenerationResult:
         ...
 
+    def analyze_images(
+        self,
+        prompt: str,
+        model: ResolvedModel,
+        *,
+        images: Sequence[MultimodalImage],
+        parameters: Mapping[str, Any] | None = None,
+    ) -> TextGenerationResult:
+        ...
+
 
 def validate_model_config(model: ResolvedModel) -> None:
     errors = []