Skip to main content

venice_ai.types.enums

Consolidated Enumerations for Venice AI Types.

This module provides all enumeration types used across the Venice AI platform, consolidating previously scattered enum definitions into a single, well-organized location. These enums provide type-safe constants for various API operations and configurations.

Available Enumerations:

  • Voice: 80+ voice options for text-to-speech generation across multiple languages
  • ResponseFormat: Audio output format options (MP3, AAC, OPUS, FLAC, WAV, PCM)
  • BillingFormatEnum: Billing data export formats (JSON, CSV)
  • ModelType: AI model type categories (TEXT, IMAGE, EMBEDDING, TTS, UPSCALE, INPAINT)

Design Philosophy:

This module serves as the single source of truth for all enumeration types, making it easy to:

  • Discover all available enum options in one place
  • Maintain consistency across the codebase
  • Avoid duplication and synchronization issues
  • Provide comprehensive documentation for all enum values

Usage Pattern:

Enums should be imported from this module rather than from legacy locations:

    from venice_ai.types.enums import Voice, ResponseFormat, BillingFormatEnum, ModelType

    # Use in API calls
    voice = Voice.AF_ALLOY
    format = ResponseFormat.MP3
    billing_format = BillingFormatEnum.JSON
    model_type = ModelType.TEXT

**Backward Compatibility:**

Legacy import paths are maintained for backward compatibility through re-exports
in the original modules (audio.py, billing.py, models.py), but new code should
import directly from this module.

Voice Objects​

class Voice(StrEnum)

Kokoro TTS voices for speech generation in the Venice AI Audio API.

This enumeration defines the Kokoro TTS voice options. For other TTS models (Qwen3, xAI, Orpheus, Inworld, Chatterbox, ElevenLabs, MiniMax, Gemini, Gradium), pass voice names as plain string literals — the voice request field is an open str, so any documented voice is accepted; see the swagger voice enum for the full list. Each Kokoro voice represents different speaker characteristics including gender, accent, and vocal qualities. Voice names follow a pattern indicating language/region and gender (e.g., af for American Female, am for American Male).

Voice Naming Convention:

  • Region Prefix: Two-letter code indicating language/locale and gender
  • Voice Name: Unique identifier for the specific voice character

Region Code Patterns:

  • af_*: American Female voices
  • am_*: American Male voices
  • bf_*: British Female voices
  • bm_*: British Male voices
  • zf_*: Chinese (Mandarin) Female voices
  • zm_*: Chinese (Mandarin) Male voices
  • ff_*: French Female voices
  • hf_*: Hindi Female voices
  • hm_*: Hindi Male voices
  • if_*: Italian Female voices
  • im_*: Italian Male voices
  • jf_*: Japanese Female voices
  • jm_*: Japanese Male voices
  • pf_*: Portuguese Female voices
  • pm_*: Portuguese Male voices
  • ef_*: Spanish Female voices
  • em_*: Spanish Male voices

Use Cases:

  • Selecting appropriate voices for different languages and content types
  • Creating region-specific or gender-specific audio content
  • Building voice selection interfaces with language and gender filters
  • Implementing voice recommendation systems based on content characteristics

Example:

::

from venice_ai.types.enums import Voice

Select American female voice

voice = Voice.AF_NOVA

Select British male voice

voice = Voice.BM_GEORGE

Select Chinese female voice

voice = Voice.ZF_XIAOXIAO

ResponseFormat Objects​

class ResponseFormat(StrEnum)

Available audio response formats for speech generation output.

This enumeration defines the supported audio file formats that can be requested when generating speech from text. The format determines the encoding, compression, and quality characteristics of the returned audio data from the text-to-speech endpoint. Different formats offer trade-offs between file size, quality, and compatibility.

Format Characteristics:

  • MP3: Lossy compression, widely compatible, good balance of quality and size
  • AAC: Advanced lossy compression, better quality than MP3 at same bitrate
  • OPUS: Modern lossy codec, excellent for speech, low latency
  • FLAC: Lossless compression, highest quality, larger file size
  • WAV: Uncompressed PCM audio, maximum quality, very large files
  • PCM: Raw uncompressed audio data, no container format

Use Cases:

  • Web applications: MP3 or AAC for broad browser compatibility
  • High-quality archival: FLAC or WAV for maximum fidelity
  • Real-time streaming: OPUS for low latency and efficient bandwidth usage
  • Audio processing: WAV or PCM for raw audio manipulation
  • Mobile applications: AAC or OPUS for efficient bandwidth usage

Quality vs. Size Trade-offs:

  • Smallest: OPUS, AAC, MP3
  • Medium: FLAC
  • Largest: WAV, PCM

Example:

::

from venice_ai.types.enums import ResponseFormat

For web playback

format = ResponseFormat.MP3

For high-quality archival

format = ResponseFormat.FLAC

For real-time streaming

format = ResponseFormat.OPUS

ResponseFormat.MP3​

MP3 = "mp3"

MP3 format - widely compatible lossy compression.

MPEG-1 Audio Layer 3 format provides good quality with moderate file sizes and universal compatibility across devices and platforms. Ideal for general-purpose web applications and content distribution.

ResponseFormat.AAC​

AAC = "aac"

AAC format - advanced lossy compression with better quality than MP3.

Advanced Audio Coding provides superior quality compared to MP3 at equivalent bitrates, with excellent compatibility on modern devices. Recommended for mobile applications and quality-focused web use.

ResponseFormat.OPUS​

OPUS = "opus"

OPUS format - modern codec optimized for speech and low latency.

Highly efficient lossy codec specifically optimized for speech and music, offering excellent quality at low bitrates with minimal latency. Perfect for real-time streaming and bandwidth-constrained applications.

ResponseFormat.FLAC​

FLAC = "flac"

FLAC format - lossless compression for maximum quality.

Free Lossless Audio Codec provides bit-perfect audio reproduction with compression, maintaining original quality while reducing file size. Ideal for archival purposes and applications requiring pristine audio.

ResponseFormat.WAV​

WAV = "wav"

WAV format - uncompressed PCM audio with container.

Waveform Audio File Format contains uncompressed PCM audio data, providing maximum compatibility and ease of processing. Suitable for audio editing and systems requiring standard uncompressed audio.

ResponseFormat.PCM​

PCM = "pcm"

PCM format - raw uncompressed audio data without container.

Pulse-Code Modulation provides raw audio samples without any container format overhead. Useful for low-level audio processing and systems requiring direct access to sample data.

BillingFormatEnum Objects​

class BillingFormatEnum(StrEnum)

Defines available output formats for billing usage data responses.

This enumeration specifies the complete set of supported data formats that can be requested when retrieving billing usage information from the Venice AI API. Each format is optimized for different use cases, from programmatic processing to data analysis and reporting, providing flexibility in how billing data is consumed and integrated into various systems and workflows.

Format Selection Strategy:

  • JSON: Ideal for programmatic access, API integrations, and web applications
  • CSV: Perfect for data analysis, spreadsheet import, and business intelligence tools

Format Details:

  • JSON format: Returns structured BillingUsageResponse with full type information, pagination metadata, and nested objects. Suitable for applications requiring precise data types, validation, and programmatic manipulation.

  • CSV format: Returns raw comma-separated values as bytes, optimized for import into Excel, Google Sheets, or data analysis tools. Includes headers and flattened data structure for easy consumption by spreadsheet applications.

Use Cases:

  • Application Integration: JSON format for real-time billing data access in applications
  • Business Analytics: CSV format for importing into Excel, Tableau, or other BI tools
  • Reporting Systems: Format selection based on downstream processing requirements
  • Data Export: CSV for sharing billing data with external systems and stakeholders
  • Automated Processing: JSON for automated billing workflows and integrations
  • Financial Reconciliation: CSV for importing into accounting systems

API Integration:

Used to specify the desired response format in billing usage API requests, allowing clients to receive data in the most appropriate format for their specific use case and processing requirements.

Example:

::

from venice_ai.types.enums import BillingFormatEnum

For programmatic access

format = BillingFormatEnum.JSON

For spreadsheet analysis

format = BillingFormatEnum.CSV

BillingFormatEnum.JSON​

JSON = "json"

JSON format - returns structured data as BillingUsageResponse.

Provides fully structured billing data with complete type information, pagination metadata, and nested objects. Ideal for programmatic access, web applications, and systems requiring precise data types and validation. The response follows the standard BillingUsageResponse model structure with full Pydantic validation and type safety.

Response Structure:

  • Type: BillingUsageResponse
  • Includes: Pagination metadata, detailed usage entries, inference details
  • Validation: Full Pydantic model validation
  • Format: Structured JSON with nested objects

BillingFormatEnum.CSV​

CSV = "csv"

CSV format - returns raw CSV data as bytes for export and analysis purposes.

Provides billing data in comma-separated values format, optimized for import into spreadsheets, business intelligence tools, and data analysis platforms. Returns raw bytes data that can be written directly to files or processed through CSV parsing libraries. Headers include all relevant billing fields for comprehensive data analysis.

Response Structure:

  • Type: bytes
  • Includes: CSV headers and flattened data rows
  • Validation: None (raw bytes)
  • Format: RFC 4180 compliant CSV with headers

CSV Columns:

Typically includes: timestamp, amount, currency, SKU, units, price per unit, notes, and inference details (when applicable).

VideoModelType Objects​

class VideoModelType(StrEnum)

Video generation model subtypes.

Distinguishes between text-to-video and image-to-video models. API-validated values from 31 video models.

VideoStatus Objects​

class VideoStatus(StrEnum)

Video generation queue status values.

Returned by the retrieve endpoint to indicate generation progress.

VideoDuration Objects​

class VideoDuration(StrEnum)

Supported video duration options.

API-validated: These are ALL durations found across 31 video models. Individual models support subsets - check model.model_spec.constraints.durations.

VideoResolution Objects​

class VideoResolution(StrEnum)

Supported video resolution options.

API-validated: These are ALL resolutions found across 31 video models. Individual models support subsets - check model.model_spec.constraints.resolutions. Some models have empty resolutions (resolution determined by model/input).

VideoPrivacy Objects​

class VideoPrivacy(StrEnum)

Video model privacy levels.

API-validated values:

  • PRIVATE: No prompt data is stored in any capacity
  • ANONYMIZED: Provider maintains anonymized prompt data

VideoAspectRatio Objects​

class VideoAspectRatio(StrEnum)

Supported video aspect ratio options.

API-validated: These are ALL aspect ratios found across 31 video models. Individual models support subsets - check model.model_spec.constraints.aspect_ratios. Image-to-video models often have empty aspect_ratios (uses input image ratio).