GET

/v1/schemas/gemini/video/gemini-omni-1.2-flash

kind inputactivity videohash 403ebb94…extracted Nov 20, 2026version currentsource upstream spec ↗
fieldrecorded source
schemaSource linked above.

generated

Source trace
a85d974f98e66ba177124a84492b44e557dac22fdf35f545c0b9c8ef8803b351
{
  "fetchedAt": 1795152603,
  "extractorVersion": "2",
  "verifiedAt": null
}

Request body

23 properties · 2 required · 88 $defs
propertytypedescriptionconstraints
model*enumThe model that generates the video.gemini-omni-1.2-flash
input*AudioContent | DocumentContent | ImageContent | Step[] | Content[] | …The input for the interaction.
generation_configGenerationConfigInput only. Configuration parameters for the model interaction.
idstringRequired. Output only. A unique identifier for the interaction completion.
statusenumRequired. Output only. The status of the interaction.in_progress | requires_action | completed | failed | cancelled | incomplete · +2 more
createdstringRequired. Output only. The time at which the response was created in ISO 8601 format (YYYY-MM-DDThh:mm:ssZ).
updatedstringRequired. Output only. The time at which the response was last updated in ISO 8601 format (YYYY-MM-DDThh:mm:ssZ).
system_instructionstringSystem instruction for the interaction.
toolsTool[]A list of tool declarations the model may call during interaction.
response_modalitiesResponseModality[]The requested modalities of the response (TEXT, IMAGE, AUDIO).deprecated
response_formatAudioResponseFormat | ImageResponseFormat | TextResponseFormat | ResponseFormat[]Enforces that the generated response is a JSON object that complies with the JSON schema specified in this field.
response_mime_typestringThe mime type of the response. This is required if response_format is set.deprecated
previous_interaction_idstringThe ID of the previous interaction, if any.
environmentobject | stringThe environment configuration for the interaction. Can be an object specifying remote environment sources or a string referencing an existing environment ID.
environment_idstringOutput only. The environment ID for the interaction. Only populated if environment config is set in the request.
service_tierenumThe service tier for the interaction.flex | standard | priority
webhook_configWebhookConfigOptional. Webhook configuration for receiving notifications when the interaction completes.
safety_settingsSafetySetting[]Safety settings for the interaction.
labelsobjectThe labels with user-defined metadata for the request.
continuation_tokenstringOpaque token to resume a long decode. Output: set when status is REQUIRES_CONTINUATION and decoding can be resumed. Input: pass the latest token back unchanged in CreateInteraction to continue decoding.format: byte
streambooleanInput only. Whether the interaction will be streamed.
storebooleanInput only. Whether to store the response and request for later retrieval.
backgroundbooleanInput only. Whether to run the model interaction in the background.

Types

88 $defs

AllowedToolsobject

The configuration for allowed tools.

modeenumThe mode of the tool choice.auto | any | none | validated
toolsstring[]The names of the allowed tools.

Annotationobject | object | object | …

Citation information for model-generated content.

AudioContentobject

An audio content block.

type*const"audio"
datastringThe audio content.format: byte
uristringThe URI of the audio.
channelsintegerThe number of audio channels.format: int32
sample_rateintegerThe sample rate of the audio.format: int32
processingAudioStaticMediaProcessing | enumHow the model processes this audio for understanding.
mime_typeenumThe mime type of the audio.audio/wav | audio/mp3 | audio/aiff | audio/aac | audio/ogg | audio/flac · +7 more

AudioMediaProcessingAudioStaticMediaProcessing

type*const"static"
start_offsetstringOptional. Segment start time, relative to the start of the audio stream. Specified as a decimal number of seconds followed by an 's' suffix, e.g., "10.5s". Must be non-negative.format: google-duration
end_offsetstringOptional. Segment end time, relative to the start of the audio stream. Specified as a decimal number of seconds followed by an 's' suffix, e.g., "30s". Must be positive and greater than `start_offset` if `start_offset` is set.format: google-duration

AudioResponseFormatobject

Configuration for audio output format.

type*const"audio"
mime_typeenumThe MIME type of the audio output.audio/mp3 | audio/ogg_opus | audio/l16 | audio/wav | audio/alaw | audio/mulaw
deliveryenumThe delivery mode for the audio output.inline | uri
sample_rateintegerSample rate in Hz.format: int32
bit_rateintegerBit rate in bits per second (bps). Only applicable for compressed formats (MP3, Opus).format: int32

AudioStaticMediaProcessingobject

type*const"static"
start_offsetstringOptional. Segment start time, relative to the start of the audio stream. Specified as a decimal number of seconds followed by an 's' suffix, e.g., "10.5s". Must be non-negative.format: google-duration
end_offsetstringOptional. Segment end time, relative to the start of the audio stream. Specified as a decimal number of seconds followed by an 's' suffix, e.g., "30s". Must be positive and greater than `start_offset` if `start_offset` is set.format: google-duration

CodeExecutionobject

A tool that can be used by the model to execute code.

type*const"code_execution"

CodeExecutionCallStepobject

Code execution call step.

arguments*CodeExecutionCallStepArgumentsRequired. The arguments to pass to the code execution.
id*stringRequired. A unique ID for this specific tool call.
type*const"code_execution_call"
signaturestringA signature hash for backend validation.format: byte

CodeExecutionCallStepArgumentsobject

The arguments to pass to the code execution.

languageenumProgramming language of the `code`.python | bash
codestringThe code to be executed.

CodeExecutionResultStepobject

Code execution result step.

result*Content[] | stringThe output of the code execution.
call_id*stringRequired. ID to match the ID from the function call block.
type*const"code_execution_result"
is_errorbooleanWhether the code execution resulted in an error.
signaturestringA signature hash for backend validation.format: byte

ComputerUseobject

A tool that can be used by the model to interact with the computer.

type*const"computer_use"
environmentenumThe environment being operated.browser | mobile | desktop
excluded_predefined_functionsstring[]The list of predefined functions that are excluded from the model call.
enable_prompt_injection_detectionbooleanWhether enable the prompt injection detection check on computer-use request.
disabled_safety_policiesenum[]Optional. Disabled safety policies for computer use.

Contentobject | object | object | …

The content of the response.

DocumentContentobject

A document content block.

type*const"document"
datastringThe document content.format: byte
uristringThe URI of the document.
resolutionenumThe resolution of the media. Applicable when PAGE_IMAGE is selected in content_layers.low | medium | high | ultra_high
processingDocumentStaticMediaProcessing | enumHow the model processes this document for understanding.
mime_typeenumThe mime type of the document.application/pdf | text/csv | application/vnd.openxmlformats-… | application/vnd.ms-powerpoint | application/vnd.oasis.opendocum… | application/vnd.openxmlformats-… · +6 more

DocumentMediaProcessingDocumentStaticMediaProcessing

type*const"static"
page_optionsPageProcessingOptionsOptions for paginated documents. Only valid for paginated document MIME types (e.g. PDF, Office documents, and presentations).
spreadsheet_optionsSpreadsheetProcessingOptionsOptions for spreadsheets. Only valid for spreadsheet MIME types (.xlsx, .xls, .ods).

DocumentStaticMediaProcessingobject

type*const"static"
page_optionsPageProcessingOptionsOptions for paginated documents. Only valid for paginated document MIME types (e.g. PDF, Office documents, and presentations).
spreadsheet_optionsSpreadsheetProcessingOptionsOptions for spreadsheets. Only valid for spreadsheet MIME types (.xlsx, .xls, .ods).

EgressRuleobject

A single domain allowlist rule with optional header injection.

domain*stringDomain to allow outbound requests to. Supports wildcards (e.g. '*.googleapis.com'). Use '*' to allow all domains.
transformobjectHeaders to inject into requests matching this rule. Key: header name (e.g., "Authorization"). Value: header value (e.g., "Bearer your-token").
credentialstringOptional. Reference to a server-managed Credential resource by ID.

EnvironmentConfigobject

Configuration for a custom environment.

type*const"remote"
sourcesSource[]
networkobject | enumNetwork configuration for the environment.
environment_idstringOptional. The environment ID for the interaction. If specified, the request will update the existing environment instead of creating a new one.
envobject | stringEnvironment variables to set in the sandbox environment.
modeenumThe compute lifecycle mode for this environment.on_demand | self_hosted
custom_configobjectOptional. Arbitrary customer-defined JSON configuration passed through to the self-hosted environment poller. Only supported when `mode` is `SELF_HOSTED`.

EnvironmentNetworkEgressAllowlistobject

Network egress configuration for the environment.

allowlistEgressRule[] | enumList of allowed domains and their configurations. Set to `"disabled"` to block all network egress.

EnvVarobject

An environment variable to set in the execution environment.

valuestringDirect string value for plain environment variables.
credentialstringOptional reference to a server-managed Credential resource by ID.

FileCitationobject

A file citation annotation.

type*const"file_citation"
document_uristringThe URI of the file.
file_namestringThe name of the file.
sourcestringSource attributed for a portion of the text.
custom_metadataobjectUser provided metadata about the retrieved context.
page_numberintegerPage number of the cited document, if applicable.format: int32
media_idstringMedia ID in-case of image citations, if applicable.
start_indexintegerStart of segment of the response that is attributed to this source.format: int32
end_indexintegerEnd of the attributed segment, exclusive.format: int32

FileSearchobject

A tool that can be used by the model to search files.

type*const"file_search"
file_search_store_namesstring[]The file search store names to search.
top_kintegerThe number of semantic retrieval chunks to retrieve.format: int32
metadata_filterstringMetadata filter to apply to the semantic retrieval documents and chunks.

FileSearchCallStepobject

File Search call step.

id*stringRequired. A unique ID for this specific tool call.
type*const"file_search_call"
signaturestringA signature hash for backend validation.format: byte

FileSearchResultStepobject

File Search result step.

call_id*stringRequired. ID to match the ID from the function call block.
type*const"file_search_result"
signaturestringA signature hash for backend validation.format: byte

Functionobject

A tool that can be used by the model.

type*const"function"
namestringThe name of the function.
descriptionstringA description of the function.
parametersanyThe JSON Schema for the function's parameters.

FunctionCallStepobject

A function tool call step.

name*stringRequired. The name of the tool to call.
arguments*objectRequired. The arguments to pass to the function.
id*stringRequired. A unique ID for this specific tool call.
type*const"function_call"

FunctionResultStepobject

Result of a function tool call.

result*ImageContent | TextContent[] | object | stringRequired. The result of the tool call.
call_id*stringRequired. ID to match the ID from the function call block.
type*const"function_result"
namestringThe name of the tool that was called.
is_errorbooleanWhether the tool call resulted in an error.

GenerationConfigobject

Configuration parameters for model interactions.

temperaturenumberControls the randomness of the output.format: float · deprecated
top_pnumberThe maximum cumulative probability of tokens to consider when sampling.format: float · deprecated
seedintegerSeed used in decoding for reproducibility.format: int32
stop_sequencesstring[]A list of character sequences that will stop output interaction.
tool_choiceobject | enumThe tool choice configuration.
thinking_levelenumThe level of thought tokens that the model should generate.minimal | low | medium | high | xhigh
thinking_summariesenumWhether to include thought summaries in the response.auto | none
max_output_tokensintegerThe maximum number of tokens to include in the response.format: int32
image_configImageConfigConfiguration for image interaction.deprecated
video_configVideoConfigConfiguration for video generation.
transcription_configTranscriptionConfigOptional. Configuration for speech recognition (transcription). If present, ASR is enabled.
structured_speech_configobject | SpeechConfig[]Speech and multi-speaker configuration.

GoogleMapsobject

A tool that can be used by the model to call Google Maps.

type*const"google_maps"
enable_widgetbooleanWhether to return a widget context token in the tool call result of the response.
latitudenumberThe latitude of the user's location.format: double
longitudenumberThe longitude of the user's location.format: double

GoogleMapsCallStepobject

Google Maps call step.

id*stringRequired. A unique ID for this specific tool call.
type*const"google_maps_call"
argumentsGoogleMapsCallStepArgumentsThe arguments to pass to the Google Maps tool.
signaturestringA signature hash for backend validation.format: byte

GoogleMapsCallStepArgumentsobject

The arguments to pass to the Google Maps tool.

queriesstring[]The queries to be executed.

GoogleMapsResultItemobject

The result of the Google Maps.

placesGoogleMapsResultPlaces[]
widget_context_tokenstring

GoogleMapsResultPlacesobject

place_idstring
namestring
urlstring
review_snippetsReviewSnippet[]

GoogleMapsResultStepobject

Google Maps result step.

result*GoogleMapsResultItem[]
call_id*stringRequired. ID to match the ID from the function call block.
type*const"google_maps_result"
signaturestringA signature hash for backend validation.format: byte

GoogleSearchobject

A tool that can be used by the model to search Google.

type*const"google_search"
search_typesenum[]The types of search grounding to enable.

GoogleSearchCallStepobject

Google Search call step.

arguments*GoogleSearchCallStepArgumentsRequired. The arguments to pass to Google Search.
id*stringRequired. A unique ID for this specific tool call.
type*const"google_search_call"
search_typeenumThe type of search grounding enabled.web_search | image_search
signaturestringA signature hash for backend validation.format: byte

GoogleSearchCallStepArgumentsobject

The arguments to pass to Google Search.

queriesstring[]Web search queries for the following-up web search.

GoogleSearchResultItemobject

The result of the Google Search.

search_suggestionsstringWeb content snippet that can be embedded in a web page or an app webview.

GoogleSearchResultStepobject

Google Search result step.

result*GoogleSearchResultItem[]Required. The results of the Google Search.
call_id*stringRequired. ID to match the ID from the function call block.
type*const"google_search_result"
is_errorbooleanWhether the Google Search resulted in an error.
signaturestringA signature hash for backend validation.format: byte

HarmCategoryenum

hate_speech | dangerous_content | harassment | sexually_explicit | jailbreak

ImageConfigobject

The configuration for image interaction.

aspect_ratiostringThe aspect ratio of the image to generate. Supported aspect ratios: 1:1, 2:3, 3:2, 3:4, 4:3, 9:16, 16:9, 21:9.
image_sizestringSpecifies the size of generated images. Supported values are `1K`, `2K`, `4K`. If not specified, the model will use default value `1K`.

ImageContentobject

An image content block.

type*const"image"
datastringThe image content.format: byte
uristringThe URI of the image.
resolutionenumThe resolution of the media.low | medium | high | ultra_high
mime_typeenumThe mime type of the image.image/png | image/jpeg | image/webp | image/heic | image/heif | image/gif · +2 more

ImageResponseFormatobject

Configuration for image output format.

type*const"image"
mime_typeenumThe MIME type of the image output.image/jpeg
deliveryenumThe delivery mode for the image output.inline | uri
aspect_ratioenumThe aspect ratio for the image output.1:1 | 2:3 | 3:2 | 3:4 | 4:3 | 4:5 · +8 more
image_sizeenumThe size of the image output.512 | 1K | 2K | 4K

InteractionsInputAudioContent | DocumentContent | ImageContent | Step[] | Content[] | …

The input for the interaction.

McpServerobject

A MCPServer is a server that can be called by the model to perform actions.

type*const"mcp_server"
namestringThe name of the MCPServer.
urlstringThe full URL for the MCPServer endpoint. Example: "https://api.example.com/mcp"
headersobjectOptional: Fields for authentication headers, timeouts, etc., if needed.
allowed_toolsAllowedTools[]The allowed tools.

McpServerToolCallStepobject

MCPServer tool call step.

name*stringRequired. The name of the tool which was called.
server_name*stringRequired. The name of the used MCP server.
arguments*objectRequired. The JSON object of arguments for the function.
id*stringRequired. A unique ID for this specific tool call.
type*const"mcp_server_tool_call"

McpServerToolResultStepobject

MCPServer tool result step.

result*ImageContent | TextContent[] | object | stringRequired. The output from the MCP server call. Can be simple text or rich content.
call_id*stringRequired. ID to match the ID from the function call block.
type*const"mcp_server_tool_result"
namestringName of the tool which is called for this specific tool call.
server_namestringThe name of the used MCP server.

MediaProcessingStaticMediaProcessing

type*const"static"
start_offsetstringOptional. Segment start time. Specified as a decimal number of seconds followed by an 's' suffix, e.g., "10.5s". Must be non-negative.format: google-duration
end_offsetstringOptional. Segment end time. Specified as a decimal number of seconds followed by an 's' suffix, e.g., "30s". Must be non-negative and greater than `start_offset` if `start_offset` is set.format: google-duration
fpsnumberOptional. Video frame-rate sampling density.format: double
content_layersenum[]Optional. Content layers to extract from the video.

MediaResolutionenum

low | medium | high | ultra_high

ModelOutputStepobject

Output generated by the model.

type*const"model_output"
contentContent[]

PageProcessingOptionsobject

Processing options for paginated documents, such as PDFs, Office documents, and presentations.

start_pageintegerOptional. 1-indexed start page, inclusive.format: int32
end_pageintegerOptional. 1-indexed end page, inclusive. Must be >= `start_page` if `start_page` is set.format: int32
content_layersenum[]Optional. Content layers to extract from a paginated document.

PlaceCitationobject

A place citation annotation.

type*const"place_citation"
place_idstringThe ID of the place, in `places/{place_id}` format.
namestringTitle of the place.
urlstringURI reference of the place.
review_snippetsReviewSnippet[]Snippets of reviews that are used to generate answers about the features of a given place in Google Maps.
start_indexintegerStart of segment of the response that is attributed to this source.format: int32
end_indexintegerEnd of the attributed segment, exclusive.format: int32

ProcessingCallStepobject

A server-initiated processing step for media analysis (e.g. video understanding).

id*stringRequired. A unique ID for this specific tool call.
type*const"processing_call"
signaturestringA signature hash for backend validation.format: byte

ProcessingResultStepobject

The result of a server-initiated media processing step.

call_id*stringRequired. ID to match the ID from the function call block.
type*const"processing_result"
signaturestringA signature hash for backend validation.format: byte

ResponseFormatobject | object | object | …

ResponseModalityenum

text | image | audio | video | document

ReviewSnippetobject

Encapsulates a snippet of a user review that answers a question about the features of a specific place in Google Maps.

titlestringTitle of the review.
urlstringA link that corresponds to the user review on Google Maps.
review_idstringThe ID of the review snippet.

SafetySettingobject

A safety setting that affects the safety-blocking behavior.

type*enumRequired. The type of harm category to be blocked.hate_speech | dangerous_content | harassment | sexually_explicit | jailbreak
threshold*enumRequired. The threshold for blocking content. If the harm probability exceeds this threshold, the content will be blocked.block_low_and_above | block_medium_and_above | block_only_high | off
methodenumOptional. The method for blocking content. If not specified, the default behavior is to use the probability score.severity | probability

ServiceTierenum

flex | standard | priority

SmartTranscriptionModeobject

Configuration for smart transcription mode.

type*const"smart"
source_textstringThe source text to be edited or rewritten via voice instructions. When provided, the model performs voice-driven editing and outputs the finalized replacement text.

Sourceobject

A source to be mounted into the environment.

typeenumgcs | inline | repository
sourcestringThe source of the environment. For Cloud Storage, this is the Cloud Storage path. For GitHub, this is the GitHub path.
targetstringWhere the source should appear in the environment.
contentstringThe inline content if `type` is `INLINE`.
encodingstringOptional encoding for inline content (e.g. `base64`).

SpeakerConfigobject

Configuration for multi-speaker and speech generation.

speakersSpeechConfig[]Individual speaker configurations.

SpeechAnnotationobject

Speech annotation for text content.

type*const"speech_metadata"
speakerstringThe speaker to associate with this turn.
stylestringStyle instruction for the speech synthesis.
start_indexintegerStart of segment of the response that is attributed to this source.format: int32
end_indexintegerEnd of the attributed segment, exclusive.format: int32

SpeechConfigobject

The configuration for speech interaction.

voicestringThe voice of the speaker.
languagestringThe language of the speech.
speakerstringThe speaker's name, it should match the speaker name given in the prompt.

SpreadsheetProcessingOptionsobject

Processing options for spreadsheets.

visible_cellsstringOptional. The range of cells to extract, in A1 notation (e.g. "A1:D10"). If not set, the full occupied area of the sheet is used.
sheet_indexintegerOptional. The 1-indexed index of the sheet to extract, in workbook order. If not set, all sheets are extracted.format: int32
content_layersenum[]Optional. Content layers to extract from a spreadsheet.

StaticMediaProcessingobject

type*const"static"
start_offsetstringOptional. Segment start time. Specified as a decimal number of seconds followed by an 's' suffix, e.g., "10.5s". Must be non-negative.format: google-duration
end_offsetstringOptional. Segment end time. Specified as a decimal number of seconds followed by an 's' suffix, e.g., "30s". Must be non-negative and greater than `start_offset` if `start_offset` is set.format: google-duration
fpsnumberOptional. Video frame-rate sampling density.format: double
content_layersenum[]Optional. Content layers to extract from the video.

Stepobject | object | object | …

A step in the interaction.

TextContentobject

A text content block.

text*stringRequired. The text content.
type*const"text"
annotationsAnnotation[]Citation information for model-generated content.

TextResponseFormatobject

Configuration for text output format.

type*const"text"
mime_typeenumThe MIME type of the text output.application/json | text/plain
schemaobjectThe JSON schema that the output should conform to. Only applicable when mime_type is application/json.

ThinkingLevelenum

minimal | low | medium | high | xhigh

ThinkingSummariesenum

auto | none

ThoughtStepobject

A thought step.

type*const"thought"
signaturestringA signature hash for backend validation.format: byte
summaryContent[]A summary of the thought.

Toolobject | object | object | …

A tool that can be used by the model.

ToolChoiceConfigobject

The tool choice configuration containing allowed tools.

allowed_toolsAllowedToolsThe allowed tools.

TranscriptionConfigobject

Configuration for speech recognition (transcription).

language_codesstring[]Optional. BCP-47 language codes providing hints about the languages present in the audio. If omitted or empty, defaults to automatic language detection.
custom_vocabularystring[]Optional. A list of custom vocabulary phrases to bias the speech recognition model toward recognizing specific terms.
modeSmartTranscriptionMode | VerbatimTranscriptionMode | enumDiscriminated transcription mode options or enum.
timestamp_granularitiesstring[]Optional. The granularity of timestamps to include in the transcription output. Supported values: "word". If empty, no timestamps are generated.deprecated
diarization_modestringOptional. Configures speaker diarization. Supported values: "speaker".deprecated
adaptation_phrasesstring[]Optional. A list of phrases to bias the ASR model towards.deprecated

TranscriptionModeobject | object

Configuration for transcription mode.

UrlCitationobject

A URL citation annotation.

type*const"url_citation"
urlstringThe URL.
titlestringThe title of the URL.
start_indexintegerStart of segment of the response that is attributed to this source.format: int32
end_indexintegerEnd of the attributed segment, exclusive.format: int32

UrlContextobject

A tool that can be used by the model to fetch URL context.

type*const"url_context"

UrlContextCallArgumentsobject

The arguments to pass to the URL context.

urlsstring[]The URLs to fetch.

UrlContextCallStepobject

URL context call step.

arguments*UrlContextCallArgumentsRequired. The arguments to pass to the URL context.
id*stringRequired. A unique ID for this specific tool call.
type*const"url_context_call"
signaturestringA signature hash for backend validation.format: byte

UrlContextResultobject

The result of the URL context.

urlstringThe URL that was fetched.
statusenumThe status of the URL retrieval.success | error | paywall | unsafe

UrlContextResultStepobject

URL context result step.

result*UrlContextResult[]Required. The results of the URL context.
call_id*stringRequired. ID to match the ID from the function call block.
type*const"url_context_result"
is_errorbooleanWhether the URL context resulted in an error.
signaturestringA signature hash for backend validation.format: byte

UserInputStepobject

Input provided by the user.

type*const"user_input"
contentContent[]

VerbatimTranscriptionModeobject

Configuration for verbatim transcription mode.

type*const"verbatim"
timestamp_granularitiesstring[]Optional. The granularity of timestamps to include in the transcription output. Supported values: "word". If empty, no timestamps are generated.
diarization_modestringOptional. Configures speaker diarization. Supported values: "speaker".

VideoConfigobject

Configuration options for video generation.

taskenumOptional task mode for video generation. If not specified, the model automatically determines the appropriate mode based on the provided text prompt and input media.text_to_video | image_to_video | reference_to_video | edit | extend

VideoContentobject

A video content block.

type*const"video"
datastringThe video content.format: byte
uristringThe URI of the video.
resolutionenumThe resolution of the media.low | medium | high | ultra_high
processingStaticMediaProcessing | enumHow the model processes this video for understanding.
namestringA user-defined name for this content block. Can be referenced by the model in the final response.
mime_typeenumThe mime type of the video.video/mp4 | video/mpeg | video/mpg | video/mov | video/avi | video/x-flv · +3 more

VideoResponseFormatobject

Configuration for video output format.

type*const"video"
deliveryenumThe delivery mode for the video output.inline | uri
aspect_ratioenumThe aspect ratio for the video output.16:9 | 9:16
durationenumThe duration for the video output, in whole seconds. gemini-omni-1.2-flash generates 3s to 15s (https://aistudio.google.com/docs/models/gemini-omni-flash.md, Output video).3s | 4s | 5s | 6s | 7s | 8s · +7 more
resolutionenumThe video output resolution. Defaults to 720p.360p | 720p | 1080p | 4k

WebhookConfigobject

Message for configuring webhook events for a request.

urisstring[]Optional. If set, these webhook URIs will be used for webhook events instead of the registered webhooks.
user_metadataobjectOptional. The user metadata that will be returned on each event emission to the webhooks.

WordInfoobject

Word-level ASR annotation for transcription output. Carries the word text, optional timing, and optional speaker attribution.

type*const"word_info"
textstringThe transcribed word.
start_offsetstringStart offset in time of the word relative to the start of the audio. Present when timestamp_granularities contains "word".format: google-duration
end_offsetstringEnd offset in time of the word relative to the start of the audio. Present when timestamp_granularities contains "word".format: google-duration
speakerstringOptional. Speaker label for this word (e.g. "spk_1", "spk_2"). Present when diarization_mode is set in TranscriptionConfig.
start_indexintegerStart of segment of the response that is attributed to this source.format: int32
end_indexintegerEnd of the attributed segment, exclusive.format: int32

Validate a payload

before you spend tokens
shell
$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"gemini","endpointId":"gemini-omni-1.2-flash","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}