hanzoai/firebase-js-sdk
2
0
Fork 0
mirror of https://github.com/hanzoai/firebase-js-sdk.git synced 2026-09-24 16:59:10 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
main
firebase-js-sdk/packages/ai/CHANGELOG.md
T
Google Open Source BotGitHubgithub-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
 02adb05c4b Version Packages (#10080) 
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2026-06-16 11:41:36 -07:00

27 KiB
Raw Permalink Blame History

@firebase/ai

2.13.1

Patch Changes

  • 9bef5ab #10035 - Deprecated GoogleMaps.enableWidget, following the announcement by the Grounding for Google Maps service.

  • 6a9d3d1 #10026 (fixes #9006) - Wrap params.systemInstruction with formatSystemInstruction() in the ChatSession constructor.

2.13.0

Minor Changes

  • f0752a0 #9800 - Add initializeDeviceModel() method to allow explicit initialization of on-device model in hybrid mode.

  • bd08cae #9926 - Change "beta" (public preview) tags to "public" (generally available) for all hybrid inference features.

Patch Changes

  • bd2644b #9956 (fixes #9938) - Fix a bug that caused ChatSession.sendMessageStream() and TemplateChatSession.sendMessageStream() to send duplicate user turns in the request.

2.12.0

Minor Changes

  • ffa39f6 #9795 - Added LiveSession.resumeSession() to allow resuming a previous LiveSession. Also added contextWindowCompression feature.

  • 86dc0db #9819 - Added support for ImageConfig (aspect ratio and size). Expanded FinishReason values to include all currently available values provided by the models.

  • 345c5f6 #9458 - AI Logic : Feature : Added support for Grounding with Google Maps.

Patch Changes

2.11.1

Patch Changes

  • 23ab5b9 #9817 - Update code execution and URL context features to reflect that they are generally available and no longer public preview.

  • 6db5af4 #9840 (fixes #9832) - Fix TemplateChatSession type so it does not error on TypeScript build. See Github issue #9832.

2.11.0

Minor Changes

Patch Changes

  • f87c15e #9791 (fixes #9792) - Fixed a bug that causes the model to error if the user specifies responseSchema or responseJsonSchema.

2.10.0

Minor Changes

Patch Changes

2.9.0

Minor Changes

  • 47f8521 #9489 - Added automatic function calling capability when using ChatSession.

Patch Changes

2.8.0

Minor Changes

  • 8123231 #9461 - AI Logic: Extend the LiveSession.receive() return types to include a LiveServiceGoingAwayNotice which signifies that the service will soon close the session.

  • f9254b6 #9475 - Remove "preview" tags from code execution and URL context documentation.

  • 2e74329 #9462 - Add params on UsageMetadata that provide cache-related usage data when the implicit caching feature is used.

Patch Changes

  • 691a506 #9469 - Internal: Add tag to log requests made to cloud while in hybrid mode.

2.7.0

Minor Changes

  • 2ecf78b #8890 (fixes #8859) - Add support for AbortSignal, allowing requests to be aborted.

  • 5579b38 #9424 - Added ability to specify thinking levels in thinkingConfig.

  • f044045 #9440 - Update default hybrid-in-cloud model to gemini-2.5-flash-lite.

2.6.1

Patch Changes

  • bc2b2cd #9399 - update the link /genai to /ailogic

  • 5c7430d #9381 - Fix generateContentStream returning wrong inferenceSource.

2.6.0

Minor Changes

2.5.0

Minor Changes

  • 22e0a1a #9291 - Deprecate sendMediaChunks() and sendMediaStream(). Instead, use the new methods added to the LiveSession class. Add sendTextRealtime(), sendAudioReatime(), and sendVideoRealtime() to the LiveSession class.

  • bc5a7c4 #9330 - Add support for audio transcriptions in the Live API.

  • c8263c4 #9315 - Add inferenceSource to the response from generateContent and generateContentStream. This property indicates whether on-device or in-cloud inference was used to generate the result.

Patch Changes

  • 44d9891 #9314 - Fix logic for merging default onDeviceParams with user-provided onDeviceParams.

2.4.0

Minor Changes

  • 0bb2fe6 #9272 - Added a sendFunctionResponses method to LiveSession, allowing function responses to be sent during realtime sessions. Fixed an issue where function responses during audio conversations caused the WebSocket connection to close. See GitHub Issue #9264.

    • Breaking Change: Changed the functionCallingHandler property in StartAudioConversationOptions so that it now must return a Promise<FunctionResponse>. This breaking change is allowed in a minor release since the Live API is in Public Preview.

  • 0ffcb26 #9254 - Added support for the URL context tool, which allows the model to access content from provided public web URLs to inform and enhance its responses.

Patch Changes

  • 2596dd1 #9255 - Imagen Generation is now Generally Available (GA).

  • 2596dd1 #9255 - The Gemini Developer API is now Generally Available (GA).

  • ea85128 #9262 - Updated SDK to handle empty parts when streaming.

  • 7a7634f #9274 - Tag code execution with beta tag (public preview).

2.3.0

Minor Changes

  • 06ab5c4 #9236 - Added a new InferenceMode option for the hybrid on-device capability: prefer_in_cloud. When this mode is selected, the SDK will attempt to use a cloud-hosted model first. If the call to the cloud-hosted model fails with a network-related error, the SDK will fall back to the on-device model, if it's available.

  • 9b8ab02 #9249 - Added Code Execution feature.

Patch Changes

  • a4848b4 #9235 - Refactor component registration.

  • c123766 #9253 - Change documentation tags for hybrid inference from "EXPERIMENTAL" to "public preview".

2.2.1

Patch Changes

2.2.0

Minor Changes

  • 984086b #9224 - Add support for the Gemini Live API.

  • 9b63cd6 #9192 - Add thoughtSummary() convenience method to EnhancedGenerateContentResponse.

  • 02280d7 #9201 - Add support for limited-use tokens with Firebase App Check. These limited-use tokens are required for an upcoming optional feature called replay protection. We recommend enabling the usage of limited-use tokens now so that when replay protection becomes available, you can enable it sooner because more of your users will be on versions of your app that send limited-use tokens.

Patch Changes

  • 84b8bed #9222 - Fixed an issue where AIError messages were too long after including an entire response body.

  • c5f08a9 #9216 - Add 'includeSafetyAttributes' field to Predict request payloads.

  • cbef6c6 #9225 - Exclude ChromeAdapterImpl code from Node entry point.

2.1.0

Minor Changes

  • e25317f #9029 - Add hybrid inference options to the Firebase AI SDK.

2.0.0

Major Changes

Minor Changes

  • a4ccd25 #9068 - Add support for Grounding with Google Search.

  • 6ab4e13 #9156 - Add support for Thinking Budget.

  • 25b60fd #9128 - Update node "engines" version to a minimum of Node 20.

Patch Changes

1.4.1

Patch Changes

1.4.0

Minor Changes

  • 1933324 #9026 - Add support for minItems and maxItems to Schema.

  • 40be2db #9047 - Add title, maximum, minimum, propertyOrdering to Schema builder

1.3.0

Minor Changes

  • e99683b #8922 - Add support for Gemini multimodal output

  • d5082f9 #8931 - Add support for the Gemini Developer API, enabling usage in a free tier, and add new AI API to accomodate new product naming.

Patch Changes

1.2.4

Patch Changes

1.2.3

Patch Changes

1.2.2

Patch Changes

1.2.1

Patch Changes

  • 648de84 #8809 - Throw an error when initializing models if appId is not defined in the given VertexAI instance.

  • faaeb48 #8832 - Label GroundingAttribution as deprecated.

1.2.0

Minor Changes

  • 25985ac #8827 - Add systemInstruction, tools, and generationConfig to CountTokensRequest.

  • 058afa2 #8741 - Added missing BlockReason and FinishReason enum values.

1.1.0

Minor Changes

  • 9d82665 #8757 - Added support for modality-based token count.

  • ce2c775 #8683 - Public Preview Added support for generating images using the Imagen 3 model.

Patch Changes

1.0.4

Patch Changes

  • 97d48c7 #8651 - FirebaseServerApp can now be initalized with an App Check token instead of invoking the App Check getToken method. This should unblock the use of App Check enforced products in SSR environments where the App Check SDK cannot be initialized.

1.0.3

Patch Changes

1.0.2

Patch Changes

  • c540ba9ee #8663 - Clear fetch timeout after request completion. Fixes an issue that caused Node scripts to hang due to a pending timeout.

1.0.1

Patch Changes

1.0.0

Major Changes

  • 479226bf3 #8475 - Release VertexAI in Firebase for general availability.

Patch Changes

0.0.4

Patch Changes

0.0.3

Patch Changes

0.0.2

Patch Changes

  • 3883133c3 #8256 - Change types paths to point to rolled-up public d.ts files. This fixes some TypeScript compiler errors users are seeing.

0.0.1

Patch Changes

Reference in New Issue View Git Blame Copy Permalink