Image input

api/openapi.yaml

@@ -318,6 +318,43 @@ paths:
         "404": { $ref: "#/components/responses/Error" }
         "409": { $ref: "#/components/responses/Error" }
 
+  /api/blobs:
+    post:
+      operationId: uploadBlob
+      description: >
+        Stores an image for use in a message. The body is the raw image; its type is detected from the bytes
+        and must be PNG, JPEG, GIF or WebP. At most 5 MiB.
+      requestBody:
+        required: true
+        content:
+          image/png: { schema: { type: string, format: binary } }
+          image/jpeg: { schema: { type: string, format: binary } }
+          image/gif: { schema: { type: string, format: binary } }
+          image/webp: { schema: { type: string, format: binary } }
+      responses:
+        "201":
+          description: The stored blob.
+          content:
+            application/json:
+              schema: { $ref: "#/components/schemas/Blob" }
+        "400": { $ref: "#/components/responses/Error" }
+        "413": { $ref: "#/components/responses/Error" }
+
+  /api/blobs/{blobId}:
+    parameters:
+      - { name: blobId, in: path, required: true, schema: { type: string } }
+    get:
+      operationId: getBlob
+      responses:
+        "200":
+          description: The blob's bytes, with its content type.
+          content:
+            image/png: { schema: { type: string, format: binary } }
+            image/jpeg: { schema: { type: string, format: binary } }
+            image/gif: { schema: { type: string, format: binary } }
+            image/webp: { schema: { type: string, format: binary } }
+        "404": { $ref: "#/components/responses/Error" }
+
   /api/inbox:
     get:
       operationId: listInbox
@@ -537,16 +574,33 @@ components:
       properties:
         prompt: { type: string }
         model: { type: string }
+        images:
+          type: array
+          items: { type: string }
+          description: Blob ids from uploadBlob, sent to the agent as images alongside the prompt.
 
     MessageCreate:
       type: object
       required: [text]
       properties:
-        text: { type: string }
+        text: { type: string, description: May be empty when images are given. }
+        images:
+          type: array
+          items: { type: string }
+          description: Blob ids from uploadBlob, sent to the agent as images alongside the text.
         model:
           type: string
           description: Model for this turn and later ones. When it differs from the session's model, the session switches before the message is sent.
 
+    Blob:
+      type: object
+      required: [id, content_type, created_by, created_at]
+      properties:
+        id: { type: string }
+        content_type: { type: string }
+        created_by: { type: string }
+        created_at: { type: string, format: date-time }
+
     Attachment:
       type: object
       required: [kind]