Indexing Revamp

README.md

@@ -610,6 +610,9 @@ response = coll.generate_text(
     model_name="pro",  # basic, pro, or ultra
     response_type="text"  # text or json
 )
+
+# Large prompts are uploaded automatically with a unique filename and
+# sent as prompt_url instead of inline JSON to avoid request payload limits.
 ```
 
 ### Video Dubbing and Translation

openapi.yaml

@@ -1915,6 +1915,14 @@ paths:
                 aspect_ratio:
                   type: string
                   example: "16:9"
+                model_name:
+                  type: string
+                  description: Use `flux` for FLUX self-inference image generation.
+                  example: "flux"
+                config:
+                  type: object
+                  description: Model configuration. For FLUX supports size, num_inference_steps, guidance_scale, negative_prompt, and seed.
+                  additionalProperties: true
                 callback_url:
                   type: string
                   example: "https://webhook.example.com/callback"
@@ -1983,16 +1991,32 @@ paths:
             schema:
               type: object
               required:
-                - prompt
                 - audio_type
               properties:
                 prompt:
                   type: string
+                  description: Prompt for music or sound effect generation.
                   example: "Generate upbeat background music"
+                text:
+                  type: string
+                  description: Text to convert to speech when `audio_type` is `voice`.
+                  example: "Hello, welcome to VideoDB."
                 audio_type:
                   type: string
-                  enum: [speech, sound_effect, music]
-                  example: "music"
+                  enum: [voice, sound_effect, music]
+                  example: "voice"
+                voice_name:
+                  type: string
+                  description: Voice name for hosted text-to-speech.
+                  example: "Default"
+                model_name:
+                  type: string
+                  description: Use `omnivoice` for OmniVoice self-inference text-to-speech.
+                  example: "omnivoice"
+                config:
+                  type: object
+                  description: Voice configuration. For OmniVoice supports instructions, ref_audio, ref_text, response_format, speed, language, and max_new_tokens.
+                  additionalProperties: true
                 callback_url:
                   type: string
                   example: "https://webhook.example.com/callback"
@@ -2004,6 +2028,39 @@ paths:
               schema:
                 $ref: '#/components/schemas/AsyncResponse'
 
+  /job/{job_id}:
+    get:
+      summary: Get self-inference generation job status
+      security:
+        - ApiKeyAuth: []
+      parameters:
+        - name: job_id
+          in: path
+          required: true
+          schema:
+            type: string
+            example: "550e8400-e29b-41d4-a716-446655440000"
+      responses:
+        '200':
+          description: Generation job status or final generated asset
+          content:
+            application/json:
+              schema:
+                type: object
+                properties:
+                  success:
+                    type: boolean
+                  status:
+                    type: string
+                    enum: [processing, done, failed]
+                  data:
+                    type: object
+                    additionalProperties: true
+                  message:
+                    type: string
+        '404':
+          description: Job not found
+
   /collection/{collection_id}/generate/text/:
     post:
       summary: Generate text using AI
@@ -2022,27 +2079,45 @@ paths:
           application/json:
             schema:
               type: object
-              required:
-                - prompt
+              description: Provide either `prompt` for inline text or `prompt_url` for a previously uploaded prompt file.
               properties:
                 prompt:
                   type: string
+                  description: Inline prompt text. Use for smaller prompts.
                   example: "Summarize the content of this video"
+                prompt_url:
+                  type: string
+                  description: URL to a previously uploaded plain-text prompt file. Use for larger prompts.
+                  example: "https://storage.googleapis.com/.../generate_text_prompt_123abc.txt"
                 video_id:
                   type: string
                   example: "m-12345"
                 model_name:
                   type: string
-                  example: "gpt-4"
+                  description: Hosted tier (`mini`, `basic`, `pro`, `ultra`) or self-hosted sandbox model name.
+                  example: "Qwen/Qwen3.5-27B"
+                sandbox_id:
+                  type: string
+                  description: Sandbox ID to use when `model_name` is a self-hosted model.
+                  example: "bx-123abc"
                 max_tokens:
                   type: integer
                   example: 500
                 temperature:
                   type: number
                   example: 0.7
+                model_config:
+                  type: object
+                  description: Additional self-hosted model configuration.
+                  additionalProperties: true
                 callback_url:
                   type: string
                   example: "https://webhook.example.com/callback"
+              oneOf:
+                - required:
+                    - prompt
+                - required:
+                    - prompt_url
       responses:
         '200':
           description: Text generation started or completed
@@ -5602,4 +5677,4 @@ tags:
   - name: Transcode
     description: Media transcoding operations
   - name: Assets
-    description: Cross-collection asset listing
+    description: Cross-collection asset listing