import sys
sys.path.insert(0, "/data/workspace/skills/venice")
from exports import (
    list_models, image_generate, image_edit, image_upscale,
    tts, transcribe, embeddings,
    list_image_styles, list_characters, list_model_traits,
    account_balance, chat_with_venice_parameters,
)

undefined

Each entry: `id, type, name, description, privacy, context_tokens, max_completion_tokens, capabilities, pricing_input_usd, pricing_output_usd, pricing_cache_input_usd, traits`.

`list_model_traits()` returns Venice's curated picks: `default`, `most_intelligent`, `most_uncensored`, `default_reasoning`, `default_vision`, `default_code`, `function_calling_default`, `fastest`. Use this when the user says "give me Venice's smartest model" — don't guess.

g = image_generate(
    "neon cyberpunk cat in the rain",
    model="venice-sd35",       # default; see list_models(type_filter='image') for others
    width=1024, height=1024,   # any aspect-ratio Venice supports
    steps=20,
    style_preset="Cinematic",  # see list_image_styles() for the 76 presets
    save_path="cat.webp",      # → output/images/cat.webp (platform convention)
)
print(g["saved_path"])

e = image_edit(
    "output/images/cat.webp",
    "make the rain heavier and add lightning",
    model="qwen-edit",         # default ($0.04/edit). Other valid IDs:
                                # firered-image-edit, grok-imagine-edit,
                                # qwen-image-2-edit, qwen-image-2-pro-edit,
                                # wan-2-7-pro-edit, flux-2-max-edit,
                                # nano-banana-pro-edit, seedream-v5-lite-edit
    save_path="cat_edited.png",
)

image_upscale("output/images/cat.webp", scale=2, save_path="cat_2x.png")

tts(
    "Welcome to Venice",
    model="tts-kokoro",        # default; alts: tts-xai-v1, tts-elevenlabs-turbo-v2-5,
                                # tts-orpheus, tts-chatterbox-hd, tts-inworld-1-5-max,
                                # tts-qwen3-0-6b, tts-qwen3-1-7b
    voice="af_alloy",          # voice list per model in Venice docs
    response_format="mp3",     # mp3 | opus | aac | flac | wav | pcm
    save_path="welcome.mp3",   # → output/audio/welcome.mp3
)

result = transcribe(
    "output/audio/welcome.mp3",
    model="openai/whisper-large-v3",  # default. Alt: stt-xai-v1
                                       # The `openai/` prefix is REQUIRED —
                                       # bare `whisper-large-v3` returns 404.
)
print(result["text"])      # transcribed text
print(result["duration"])  # seconds

out = embeddings(["hello world", "second sentence"])

undefined

list_characters(limit=20)  # [{slug, name, description, tags}, ...]

undefined

chat_with_venice_parameters(
    "What's the latest Bitcoin price?",
    venice_parameters={
        "enable_web_search": "on",            # "auto" | "on" | "off"
        "include_venice_system_prompt": False, # drop Venice's default sysprompt
        "enable_web_citations": True,         # ask for inline citations
    },
)

from exports import video_generate, video_quote, list_models

`video_generate` accepts every queue parameter via `**queue_kwargs`:

| Field | Purpose |
|---|---|
| `negative_prompt` | What to avoid |
| `image_url` | First-frame reference (image-to-video) |
| `end_image_url` | Last-frame reference (transition) |
| `audio_url` | Background music input (WAV/MP3, ≤30s, ≤15MB) |
| `video_url` | Video-to-video / upscale input (MP4/MOV/WebM) |
| `reference_image_urls` | Up to 9 character/style references |
| `elements` | Up to 4 advanced elements (Kling O3 R2V style); reference in prompt as `@Element1` |
| `scene_image_urls` | Up to 4 scene references; reference as `@Image1` |
| `upscale_factor` | 1 / 2 / 4 — for upscale models (use instead of resolution) |
| `delete_media_on_completion` | Auto-delete from Venice storage after retrieve |

**Manual loop** when you want fine control (interactive ETA, custom storage, batched jobs):

```python
queued = video_queue(model="...", prompt="...", duration="5s",
                      resolution="720p", aspect_ratio="16:9")

**Video transcription (YouTube only)**:

```python
video_transcribe_youtube("https://www.youtube.com/watch?v=...")

For arbitrary local audio/video files, use `transcribe()` (uploads to /audio/transcriptions and accepts file paths). For non-YouTube hosted video, strip audio with ffmpeg first then call `transcribe()`.

**Video errors** beyond the table above:

| Code | Meaning |
|---|---|
| 400 | Bad params: model doesn't support that duration/resolution combo, missing `image_url` for i2v, or `prompt` empty |
| 402 | Insufficient balance — top up at venice.ai |
| 413 | Payload too big — use hosted URLs instead of base64 data URIs |
| 422 | Content policy violation (rare on Venice but possible on i2v) |
| 503 | Queue saturated — wait and retry |

Gotchas:
- `duration` is required even for `Auto` (pass it explicitly).
- `download_url` from queue is valid 24h — fetch promptly.
- Upscale models require `upscale_factor`, NOT `resolution`.
- Quote varies wildly by model and duration — wan 5s @ 720p ≈ $0.55, seedance-2-0-pro 10s @ 1080p can be $5+.