feat:Adopt GA session schemas; remove event schemas; update OpenAPI refs #223
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
WalkthroughReplaced prerelease session schemas with GA variants in the OpenAPI spec, removed two public event schemas, adjusted anyOf references to GA types, removed one client event type, tweaked descriptions (including idle turn-detection notes), and dropped a const metadata flag. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor Client
participant API as Realtime API
rect rgb(235, 245, 255)
note over Client,API: GA session creation flow
Client->>API: POST /sessions (create)
alt Standard session
API-->>Client: 201 RealtimeSessionCreateResponseGA
else Transcription session
API-->>Client: 201 RealtimeTranscriptionSessionCreateResponseGA
end
end
rect rgb(245, 235, 255)
note over Client,API: Session lifecycle events
Client->>API: WebSocket subscribe
API-->>Client: ...updates (e.g., TranscriptionSessionUpdated)
note right of API: "Created" event removed from public stream
opt Idle turn detection
API-->>Client: timeout_triggered (on idle threshold)
end
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Pre-merge checks (2 passed, 1 warning)❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
Poem
✨ Finishing Touches🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
coderabbitai
bot
changed the title
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 5
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/libs/tryAGI.OpenAI/openapi.yaml (1)
24582-24590: GA session schema needs a concretetype(and likelyobject) to satisfy discriminator and const semantics.
required: [client_secret, type]is declared, buttypeisn’t defined inproperties(as shown). Add a constanttype: "realtime"; consider also const-encodingobject: "realtime.session".Apply:
RealtimeSessionCreateResponseGA: required: - client_secret - type type: object properties: + type: + type: string + enum: [realtime] + description: The type of session. Always `realtime` for realtime sessions. + x-stainless-const: true + object: + type: string + enum: [realtime.session] + description: The object type. Always `realtime.session`. + x-stainless-const: true audio:Also verify
client_secretis actually returned by this endpoint; if not, drop it fromrequired.
🧹 Nitpick comments (1)
src/libs/tryAGI.OpenAI/openapi.yaml (1)
24450-24582: Non-GA and GA session objects coexist; consider deprecating or removing the old one.If nothing references
RealtimeSessionCreateResponseanymore, mark it deprecated or remove it to avoid confusion and accidental use.I can search the repo to confirm references before you remove it.
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (79)
src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI..JsonSerializerContext.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeClientEvent.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAAudioInputTurnDetectionEagerness.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAAudioInputTurnDetectionEagernessNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAAudioInputTurnDetectionType.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAAudioInputTurnDetectionTypeNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAIncludeItem.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAIncludeItemNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAMaxOutputTokens.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAMaxOutputTokensNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAModel.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAModelNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAOutputModalitie.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAOutputModalitieNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGATracingEnum.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGATracingEnumNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGAType.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeSessionCreateResponseGATypeNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeTranscriptionSessionCreateResponseGAIncludeItem.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeTranscriptionSessionCreateResponseGAIncludeItemNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeTranscriptionSessionCreateResponseGAType.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.RealtimeTranscriptionSessionCreateResponseGATypeNullable.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonConverters.Session2.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.JsonSerializerContextTypes.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeClientEvent.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeClientEventSessionUpdate.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeServerEventTranscriptionSessionCreated.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateRequestGAAudioInputTurnDetection.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponse.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseAudio.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseAudioInput.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseAudioInputNoiseReduction.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseAudioInputTurnDetection.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseAudioOutput.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGA.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGA.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudio.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudio.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioInput.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioInput.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioInputNoiseReduction.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioInputNoiseReduction.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioInputTurnDetection.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioInputTurnDetection.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioInputTurnDetectionEagerness.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioInputTurnDetectionType.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioOutput.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAAudioOutput.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAClientSecret.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAClientSecret.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAIncludeItem.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAMaxOutputTokens.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAModel.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAOutputModalitie.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGATracingEnum.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGATracingEnum2.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGATracingEnum2.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGATracingEnumMetadata.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGATracingEnumMetadata.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseGAType.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseOutputModalities.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseOutputModalities.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseTracingEnum2.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseTracingEnumMetadata.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseTurnDetection.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeSessionCreateResponseTurnDetection.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGA.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGA.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAAudio.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAAudio.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAAudioInput.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAAudioInput.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAAudioInputNoiseReduction.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAAudioInputNoiseReduction.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAAudioInputTurnDetection.Json.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAAudioInputTurnDetection.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAIncludeItem.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.RealtimeTranscriptionSessionCreateResponseGAType.g.csis excluded by!**/generated/**src/libs/tryAGI.OpenAI/Generated/tryAGI.OpenAI.Models.Session2.g.csis excluded by!**/generated/**
📒 Files selected for processing (1)
src/libs/tryAGI.OpenAI/openapi.yaml(6 hunks)
🔇 Additional comments (3)
src/libs/tryAGI.OpenAI/openapi.yaml (3)
24319-24319: New event mention: verifytimeout_triggeredis defined.You now document an emitted
timeout_triggeredevent. Ensure a corresponding event schema/name exists elsewhere, or rename here to match the actual event.Would you like me to search and list where this event is (or isn’t) defined across the spec?
24622-24622: Repeat of new event mention: verifytimeout_triggeredexists.Same concern as earlier idle-timeout description: ensure the event is defined and consistently named.
24955-24957: ValidateRealtimeAudioFormatssupports string literalpcm16used in example.The example uses
"format": "pcm16", while other examples show an object form. Please confirm#/components/schemas/RealtimeAudioFormatspermits the string variant and includespcm16.If not, align the example or widen the schema (e.g.,
oneOfstring enum | object form).Also applies to: 25006-25006
| description: 'The event type, must be `session.update`.' | ||
| x-stainless-const: true | ||
| description: "Send this event to update the session’s configuration.\nThe client may send this event at any time to update any field\nexcept for `voice` and `model`. `voice` can be updated only if there have been no other\naudio outputs yet. \n\nWhen the server receives a `session.update`, it will respond\nwith a `session.updated` event showing the full, effective configuration.\nOnly the fields that are present in the `session.update` are updated. To clear a field like\n`instructions`, pass an empty string. To clear a field like `tools`, pass an empty array.\nTo clear a field like `turn_detection`, pass `null`.\n" | ||
| description: "Send this event to update the session’s configuration.\nThe client may send this event at any time to update any field\nexcept for `voice` and `model`. `voice` can be updated only if there have been no other audio outputs yet.\n\nWhen the server receives a `session.update`, it will respond\nwith a `session.updated` event showing the full, effective configuration.\nOnly the fields that are present in the `session.update` are updated. To clear a field like\n`instructions`, pass an empty string. To clear a field like `tools`, pass an empty array.\nTo clear a field like `turn_detection`, pass `null`.\n" |
There was a problem hiding this comment.
Doc says you can clear turn_detection with null, but schema doesn’t allow null.
The description instructs clients to send turn_detection: null, yet the schema blocks nulls (no nullable: true). This will break clients/codegen.
Follow-up: See suggested fixes on Lines 24561-24576 and 24966-24981 to mark these fields nullable.
🤖 Prompt for AI Agents
In src/libs/tryAGI.OpenAI/openapi.yaml around line 21728, the documentation
tells clients to send turn_detection: null but the schema does not permit nulls;
update the OpenAPI schema by adding nullable: true (or using type: ['object',
'null'] where appropriate) to the turn_detection property definitions referenced
in the file and to the other occurrences indicated around lines 24561-24576 and
24966-24981 so the field accepts null; ensure any component/schema definitions
and request/response examples remain consistent with the nullable change.
Comment on lines
+22211
to
22215
| - $ref: '#/components/schemas/RealtimeSessionCreateResponseGA' | ||
| - $ref: '#/components/schemas/RealtimeTranscriptionSessionCreateResponseGA' | ||
| description: "The session configuration for either a realtime or transcription session.\n" | ||
| discriminator: | ||
| propertyName: type |
There was a problem hiding this comment.
Discriminator likely broken: GA “session” schema lacks a concrete type value + no mapping.
discriminator.propertyName: type is set, but RealtimeSessionCreateResponseGA appears to not define a type property/value, while the transcription GA variant does. Add a constant type: "realtime" and explicit discriminator mapping.
Apply:
discriminator:
propertyName: type
+ mapping:
+ realtime: '#/components/schemas/RealtimeSessionCreateResponseGA'
+ transcription: '#/components/schemas/RealtimeTranscriptionSessionCreateResponseGA'And add type in the GA session schema (see Lines 24582-24590 comment).
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| - $ref: '#/components/schemas/RealtimeSessionCreateResponseGA' | |
| - $ref: '#/components/schemas/RealtimeTranscriptionSessionCreateResponseGA' | |
| description: "The session configuration for either a realtime or transcription session.\n" | |
| discriminator: | |
| propertyName: type | |
| discriminator: | |
| propertyName: type | |
| mapping: | |
| realtime: '#/components/schemas/RealtimeSessionCreateResponseGA' | |
| transcription: '#/components/schemas/RealtimeTranscriptionSessionCreateResponseGA' |
🤖 Prompt for AI Agents
In src/libs/tryAGI.OpenAI/openapi.yaml around lines 22211-22215 (and update the
GA session schema at lines ~24582-24590), the discriminator is misconfigured:
discriminator.propertyName is set to "type" but RealtimeSessionCreateResponseGA
does not declare a constant type value and there's no mapping entry. Fix by
adding a required, constant property type: "realtime" to the
RealtimeSessionCreateResponseGA schema, and add an explicit
discriminator.mapping entry mapping the discriminator value "realtime" to the
RealtimeSessionCreateResponseGA component (e.g.,
"#/components/schemas/RealtimeSessionCreateResponseGA") so the OAS discriminator
can resolve the subtype correctly.
Comment on lines
+24561
to
+24576
| turn_detection: | ||
| type: object | ||
| properties: | ||
| prefix_padding_ms: | ||
| type: integer | ||
| description: "Amount of audio to include before the VAD detected speech (in\nmilliseconds). Defaults to 300ms.\n" | ||
| silence_duration_ms: | ||
| type: integer | ||
| description: "Duration of silence to detect speech stop (in milliseconds). Defaults\nto 500ms. With shorter values the model will respond more quickly,\nbut may jump in on short pauses from the user.\n" | ||
| threshold: | ||
| type: number | ||
| description: "Activation threshold for VAD (0.0 to 1.0), this defaults to 0.5. A\nhigher threshold will require louder audio to activate the model, and\nthus might perform better in noisy environments.\n" | ||
| type: | ||
| type: string | ||
| description: "Type of turn detection, only `server_vad` is currently supported.\n" | ||
| description: "Configuration for turn detection. Can be set to `null` to turn off. Server\nVAD means that the model will detect the start and end of speech based on\naudio volume and respond at the end of user speech.\n" |
There was a problem hiding this comment.
Schema-Doc mismatch: allow null for turn_detection and add bounds for threshold.
- Description: “Can be set to
nullto turn off.” Butnullableisn’t set. thresholdis described as 0.0–1.0; encode that as min/max for better validation.
Apply:
turn_detection:
type: object
+ nullable: true
properties:
prefix_padding_ms:
type: integer
silence_duration_ms:
type: integer
threshold:
type: number
+ minimum: 0
+ maximum: 1
type:
type: string📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| turn_detection: | |
| type: object | |
| properties: | |
| prefix_padding_ms: | |
| type: integer | |
| description: "Amount of audio to include before the VAD detected speech (in\nmilliseconds). Defaults to 300ms.\n" | |
| silence_duration_ms: | |
| type: integer | |
| description: "Duration of silence to detect speech stop (in milliseconds). Defaults\nto 500ms. With shorter values the model will respond more quickly,\nbut may jump in on short pauses from the user.\n" | |
| threshold: | |
| type: number | |
| description: "Activation threshold for VAD (0.0 to 1.0), this defaults to 0.5. A\nhigher threshold will require louder audio to activate the model, and\nthus might perform better in noisy environments.\n" | |
| type: | |
| type: string | |
| description: "Type of turn detection, only `server_vad` is currently supported.\n" | |
| description: "Configuration for turn detection. Can be set to `null` to turn off. Server\nVAD means that the model will detect the start and end of speech based on\naudio volume and respond at the end of user speech.\n" | |
| turn_detection: | |
| type: object | |
| nullable: true | |
| properties: | |
| prefix_padding_ms: | |
| type: integer | |
| silence_duration_ms: | |
| type: integer | |
| threshold: | |
| type: number | |
| minimum: 0 | |
| maximum: 1 | |
| type: | |
| type: string |
🤖 Prompt for AI Agents
In src/libs/tryAGI.OpenAI/openapi.yaml around lines 24561 to 24576, the schema
for turn_detection incorrectly omits nullable and the threshold lacks numeric
bounds; update the turn_detection property to include nullable: true (or
explicitly allow ["object","null"] if using oneOf/anyOf style) so it can be set
to null as described, and add minimum: 0.0 and maximum: 1.0 to the threshold
property (ensuring type: number/float) to enforce the documented 0.0–1.0 range.
| description: "Configuration for turn detection. Can be set to `null` to turn off. Server\nVAD means that the model will detect the start and end of speech based on\naudio volume and respond at the end of user speech.\n" | ||
| description: "A Realtime session configuration object.\n" | ||
| x-oaiMeta: | ||
| example: "{\n \"id\": \"sess_001\",\n \"object\": \"realtime.session\",\n \"expires_at\": 1742188264,\n \"model\": \"gpt-realtime\",\n \"output_modalities\": [\"audio\"],\n \"instructions\": \"You are a friendly assistant.\",\n \"tools\": [],\n \"tool_choice\": \"none\",\n \"max_output_tokens\": \"inf\",\n \"tracing\": \"auto\",\n \"truncation\": \"auto\",\n \"prompt\": null,\n \"audio\": {\n \"input\": {\n \"format\": {\n \"type\": \"audio/pcm\",\n \"rate\": 24000\n },\n \"transcription\": { \"model\": \"whisper-1\" },\n \"noise_reduction\": null,\n \"turn_detection\": null\n },\n \"output\": {\n \"format\": {\n \"type\": \"audio/pcm\",\n \"rate\": 24000\n },\n \"voice\": \"alloy\",\n \"speed\": 1.0\n }\n }\n}\n" |
There was a problem hiding this comment.
Example contains fields not in schema (truncation, prompt).
The example JSON shows truncation and prompt which aren’t defined in this schema block. This will confuse users and can fail example validation.
Apply:
- "truncation": "auto",
- "prompt": null,(Or add these fields to the schema if they are intended.)
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| example: "{\n \"id\": \"sess_001\",\n \"object\": \"realtime.session\",\n \"expires_at\": 1742188264,\n \"model\": \"gpt-realtime\",\n \"output_modalities\": [\"audio\"],\n \"instructions\": \"You are a friendly assistant.\",\n \"tools\": [],\n \"tool_choice\": \"none\",\n \"max_output_tokens\": \"inf\",\n \"tracing\": \"auto\",\n \"truncation\": \"auto\",\n \"prompt\": null,\n \"audio\": {\n \"input\": {\n \"format\": {\n \"type\": \"audio/pcm\",\n \"rate\": 24000\n },\n \"transcription\": { \"model\": \"whisper-1\" },\n \"noise_reduction\": null,\n \"turn_detection\": null\n },\n \"output\": {\n \"format\": {\n \"type\": \"audio/pcm\",\n \"rate\": 24000\n },\n \"voice\": \"alloy\",\n \"speed\": 1.0\n }\n }\n}\n" | |
| - "truncation": "auto", | |
| - "prompt": null, |
🤖 Prompt for AI Agents
In src/libs/tryAGI.OpenAI/openapi.yaml around line 24579, the example JSON
includes fields "truncation" and "prompt" that are not defined in the
corresponding schema; remove those fields from the example to match the schema,
or if they are intended properties, add "truncation" and/or "prompt" definitions
to the schema (with appropriate types and descriptions) and update any required
lists so the example validates against the schema.
Comment on lines
+24966
to
+24981
| type: object | ||
| properties: | ||
| prefix_padding_ms: | ||
| type: integer | ||
| description: "Amount of audio to include before the VAD detected speech (in\nmilliseconds). Defaults to 300ms.\n" | ||
| silence_duration_ms: | ||
| type: integer | ||
| description: "Duration of silence to detect speech stop (in milliseconds). Defaults\nto 500ms. With shorter values the model will respond more quickly,\nbut may jump in on short pauses from the user.\n" | ||
| threshold: | ||
| type: number | ||
| description: "Activation threshold for VAD (0.0 to 1.0), this defaults to 0.5. A\nhigher threshold will require louder audio to activate the model, and\nthus might perform better in noisy environments.\n" | ||
| type: | ||
| type: string | ||
| description: "Type of turn detection, only `server_vad` is currently supported.\n" | ||
| description: "Configuration for turn detection. Can be set to `null` to turn off. Server\nVAD means that the model will detect the start and end of speech based on\naudio volume and respond at the end of user speech.\n" | ||
| description: "Configuration for input audio for the session.\n" |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Transcription GA: allow null for turn_detection and add bounds for threshold.
Mirror the realtime session changes: set nullable: true and encode 0–1 range.
Apply:
turn_detection:
type: object
+ nullable: true
properties:
prefix_padding_ms:
type: integer
silence_duration_ms:
type: integer
threshold:
type: number
+ minimum: 0
+ maximum: 1
type:
type: string📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| type: object | |
| properties: | |
| prefix_padding_ms: | |
| type: integer | |
| description: "Amount of audio to include before the VAD detected speech (in\nmilliseconds). Defaults to 300ms.\n" | |
| silence_duration_ms: | |
| type: integer | |
| description: "Duration of silence to detect speech stop (in milliseconds). Defaults\nto 500ms. With shorter values the model will respond more quickly,\nbut may jump in on short pauses from the user.\n" | |
| threshold: | |
| type: number | |
| description: "Activation threshold for VAD (0.0 to 1.0), this defaults to 0.5. A\nhigher threshold will require louder audio to activate the model, and\nthus might perform better in noisy environments.\n" | |
| type: | |
| type: string | |
| description: "Type of turn detection, only `server_vad` is currently supported.\n" | |
| description: "Configuration for turn detection. Can be set to `null` to turn off. Server\nVAD means that the model will detect the start and end of speech based on\naudio volume and respond at the end of user speech.\n" | |
| description: "Configuration for input audio for the session.\n" | |
| turn_detection: | |
| type: object | |
| nullable: true | |
| properties: | |
| prefix_padding_ms: | |
| type: integer | |
| silence_duration_ms: | |
| type: integer | |
| threshold: | |
| type: number | |
| minimum: 0 | |
| maximum: 1 | |
| type: | |
| type: string |
🤖 Prompt for AI Agents
In src/libs/tryAGI.OpenAI/openapi.yaml around lines 24966 to 24981, the
transcription GA schema needs to allow null for the turn_detection field and
enforce a 0–1 range for the threshold; change the turn_detection property to
include nullable: true and add minimum: 0 and maximum: 1 (or inclusive bounds
appropriate to your schema flavor) under the threshold property so the OpenAPI
spec mirrors the realtime session behavior.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary by CodeRabbit