Generate Story (Codex)

You will be given:

• A story subject (prompt) for James.
• A folder path that may contain reference images and may include a date in the folder name.
• Optionally a markdown file name that may include a date.

Required environment variables:

• REPLICATE_API_TOKEN for Replicate.
• STORY_API_TOKEN for the worker automation API.
• STORY_API_BASE_URL (optional). If not set, use https://bedtimestories.bruce-hart.workers.dev.

Use the following workflow exactly.

Step 0: Resolve story date

1. If the folder name contains a YYYY-MM-DD date, use that.
2. Else if the markdown file name contains a YYYY-MM-DD date, use that.
3. Else use the current date in YYYY-MM-DD.

Step 1: Write the story text (Markdown)

Follow these instructions exactly:

For the given prompt, write a bedtime story including a title for my six year old son James. Make the text simple and phonics friendly for a beginning reader. Return the title separately and do not include it in the story content.

[Story Content]

In addition to Mom and Dad in his family, James has a three year old sister named Grace. He also has a small tuxedo cat named Trixie. In his family is also Granny and Grandpa Rick and Grandma and Grandpa Bruce. Only include people that are relevant to the story. Do not include people just to include them.

Do not use words in any image generated. Images should be in landscape format and have a cartoon aesthetic. Only include people in the images that are relevant to the scene of the story being visualized.

James has fair skin and short brown hair.

Mom has fair skin, shoulder length brown hair with blonde highlights and glasses.

Grace has long brown hair and fair skin.

Dad is bald and clean shaven with brown hair and thick dark eyebrows.

Granny is short with short gray hair.

Grandma has short blonde hair and glasses.

Grandpa is bald and clean shaven.

Trixie is a black cat with short legs, white paws, black chin, black face and a white chest.

Story format requirements:

• Plain text only (no Markdown).
• Do not include the title in the story content.
• Use short paragraphs separated by blank lines.
• Keep sentences short and easy to read.

Step 2: Generate a cover image (Replicate)

Model: google/nano-banana-pro

Image requirements:

• Landscape, 16:9.
• Cartoon aesthetic.
• No text, letters, or signage.
• Only include characters relevant to the selected scene.
• Incorporate reference images as guidance when available.

Convert local reference images to data URIs for image_input (list up to 14). Example (Python):

python - <<'PY'
import base64, json, mimetypes, pathlib
paths = list(pathlib.Path("PATH_TO_FOLDER").glob("*.*"))
inputs = []
for p in paths:
    mime = mimetypes.guess_type(p.name)[0] or "application/octet-stream"
    data = base64.b64encode(p.read_bytes()).decode("ascii")
    inputs.append(f"data:{mime};base64,{data}")
print(json.dumps(inputs))
PY

If there are no reference images, set image_input to an empty list.

Create the image prediction:

curl -s https://api.replicate.com/v1/models/google/nano-banana-pro/predictions \
  -H "Authorization: Token $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "input": {
      "prompt": "YOUR_IMAGE_PROMPT",
      "image_input": [DATA_URI_LIST],
      "aspect_ratio": "16:9",
      "resolution": "2K",
      "output_format": "jpg",
      "safety_filter_level": "block_only_high"
    }
  }'

Poll until status is succeeded:

curl -s https://api.replicate.com/v1/predictions/PREDICTION_ID \
  -H "Authorization: Token $REPLICATE_API_TOKEN"

Save the output URL when ready.

Alternate image model (use if nano-banana-pro fails):

• Model: black-forest-labs/flux-1.1-pro
• Aspect ratio: 3:2
• Output format: png
• Schema: https://replicate.com/black-forest-labs/flux-1.1-pro/api/schema

Create the Flux image prediction:

curl -s https://api.replicate.com/v1/models/black-forest-labs/flux-1.1-pro/predictions \
  -H "Authorization: Token $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "input": {
      "prompt": "YOUR_IMAGE_PROMPT",
      "aspect_ratio": "3:2",
      "output_format": "png"
    }
  }'

Step 3: Generate a short video (Replicate)

Primary model: google/veo-3-fast

Video requirements:

• 16:9 landscape.
• Cartoon aesthetic.
• No text or letters.
• Use the generated image as the image input for consistency.
• Use the story to build a concise scene prompt.

Create the video prediction:

curl -s https://api.replicate.com/v1/models/google/veo-3-fast/predictions \
  -H "Authorization: Token $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "input": {
      "prompt": "YOUR_VIDEO_PROMPT",
      "image": "GENERATED_IMAGE_URL",
      "aspect_ratio": "16:9",
      "duration": 4,
      "resolution": "1080p",
      "generate_audio": true
    }
  }'

Poll until status is succeeded and record the output URL.

If Veo fails (status failed or sensitive content error), fall back to PixVerse:

• Model: pixverse/pixverse-v5
• Use 5 seconds, normal (effect None), 540p.

Create the PixVerse video prediction:

curl -s https://api.replicate.com/v1/models/pixverse/pixverse-v5/predictions \
  -H "Authorization: Token $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "input": {
      "prompt": "YOUR_VIDEO_PROMPT",
      "image": "GENERATED_IMAGE_URL",
      "aspect_ratio": "16:9",
      "duration": 8,
      "quality": "540p",
      "effect": "None"
    }
  }'

Poll until status is succeeded and record the output URL.

Download the generated assets locally before uploading:

curl -L "IMAGE_OUTPUT_URL" -o /tmp/story-image.jpg
curl -L "VIDEO_OUTPUT_URL" -o /tmp/story-video.mp4

Step 4: Upload media to R2 via worker API

Base URL:

• If STORY_API_BASE_URL is set, use that.
• Else use https://bedtimestories.bruce-hart.workers.dev.

Upload image:

curl -s "$STORY_API_BASE_URL/api/media" \
  -H "X-Story-Token: $STORY_API_TOKEN" \
  -F "file=@/path/to/image.jpg"

Upload video:

curl -s "$STORY_API_BASE_URL/api/media" \
  -H "X-Story-Token: $STORY_API_TOKEN" \
  -F "file=@/path/to/video.mp4"

Each response returns { "key": "..." }. Keep those keys.

Step 5: Create the story record

Send the story to the worker:

curl -s "$STORY_API_BASE_URL/api/stories" \
  -H "X-Story-Token: $STORY_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "TITLE",
    "content": "MARKDOWN_STORY",
    "date": "YYYY-MM-DD",
    "image_url": "IMAGE_KEY",
    "video_url": "VIDEO_KEY"
  }'

Final output to user

Return:

• Title
• Story content (plain text)
• Image key
• Video key
• Story id