Skip to content

Latest commit

 

History

History
1211 lines (697 loc) · 83.1 KB

CHANGELOG.md

File metadata and controls

1211 lines (697 loc) · 83.1 KB

Changelog

1.79.0 (2026-11-18)

Features

  • sdk: Add tool retrieval and function result schemas in public SDK (25bf4e8)

Bug Fixes

  • send SDK User-Agent and x-goog-api-client headers on the Live WebSocket handshake (99f48a5), refs #1045

1.78.0 (2026-11-17)

⚠ BREAKING CHANGES

  • Default Gemini Developer API version is now v1 (was v1beta) (402dc81)
    • Clients targeting the Gemini Developer API (the default backend, genai.BackendGeminiAPI) now default to API version v1 instead of v1beta.

      Note for Go users: Although this default change is a runtime breaking change, we are shipping it in a minor v1.x release rather than v2.0.0 as the less disruptive of the two options. In Go, a v2 major version bump changes the module path to google.golang.org/genai/v2, which breaks every user and downstream library by forcing import rewrites across the ecosystem—even though no Go types or function signatures changed. Shipping in v1.x keeps existing imports and builds working, leaves most code unaffected, and lets anyone who needs the previous v1beta behavior restore it with a single config field or environment variable (see below).

    • No effect on Gemini Enterprise (Vertex AI) defaults: Clients using genai.BackendVertexAI (or GOOGLE_GENAI_USE_VERTEXAI=true) are unaffected by this default change and continue to default to v1beta1.
    • Some models are not available on v1, notably: the gemini-flash-latest, gemini-flash-lite-latest, and gemini-pro-latest aliases, some -preview models, and models that depend on v1beta-only features (such as TTS models). They exist only on v1beta. To keep using them, use a specific stable model version, or switch back to v1beta as shown below.
    • Some APIs are not available on v1: The Live API (client.Live), context caching (client.Caches), and beta platform services (client.Agents, client.Decisions, client.Webhooks) are served only on v1beta in the Gemini Developer API. To keep using them, switch back to v1beta as shown below.
    • How to configure or restore the previous default via environment variable: Set GOOGLE_GENAI_API_VERSION (works for both Gemini Developer API and Enterprise / Vertex AI clients when no version is set in code):
      # Restores the beta default ('v1beta' on Gemini Developer API, 'v1beta1' on Enterprise / Vertex AI)
      export GOOGLE_GENAI_API_VERSION=beta
      
      # Or set an explicit API version string on either platform (e.g. 'v1', 'v1beta', 'v1beta1')
      export GOOGLE_GENAI_API_VERSION=v1beta
    • How to set the API version in code: Set APIVersion in HTTPOptions (takes precedence over GOOGLE_GENAI_API_VERSION):
      import (
      	"context"
      
      	"google.golang.org/genai"
      )
      
      client, err := genai.NewClient(context.Background(), &genai.ClientConfig{
      	HTTPOptions: genai.HTTPOptions{APIVersion: "v1beta"},
      })

Features

  • add gemini-omni-1.2-flash and make it the gemini-api video default (04c227d)
  • Add proactive_video to ProactivityConfig for the Live API. (4ac9851)
  • Add source_text to AudioTranscriptionConfig and SmartTranscriptionMode. (2b4e897)
  • Add XHIGH and xhigh to ThinkingLevel across GenAI SDKs and CLI. (78c23d6)
  • add xhigh thinking level to Interactions ThinkingLevel (5a3af4c)
  • Add MALFORMED_RESPONSE to FinishReason (6ba4b48), refs #2937
  • Add missing enum values to FinishReason and ReinforcementTuningThinkingLevel (9517de6)
  • add the Decisions API (client.decisions.create) to the GenAI SDKs (4e1c734)
  • Add tool search types to the Interactions SDKs (f2b5eea)
  • add usage_metadata (including traffic_type) and truncated to EmbedContentResponse for Gemini Enterprise API (850cdca)
  • cli: use gemini-3.5-transcribe as default model for transcribe command (968e135)
  • default Gemini API version to v1 instead of v1beta across SDKs (402dc81)
  • Disable GenerateVideos for Gemini API, update tests/READMEs (9cf04de)
  • environments: introduce self-hosted BYOS mode, system_instruction, custom_config, and environment_key in SDKs and CLI (eeafe6c)
  • environments: introduce self-hosted BYOS mode, system_instruction, custom_config, and environment_key in SDKs and CLI (ac12b34)
  • interactions: add gemini-4-carbon and gemini-omni-1.2-flash to Model enum (e08c6be)
  • interactions: add gemini-4-hydrogen to Model enum (1b7b34c)
  • interactions: add Office and OpenDocument MIME types (36ba34e)
  • interactions: expose static file processing options (5d76cf0)
  • interactions: support multimodal content in CodeExecutionResultStep.result (0872815)
  • support batch inputs and usage metadata for Gemini embedding models on Gemini Enterprise API (c1a7aa2)
  • support GOOGLE_GENAI_API_VERSION environment variable across SDKs (402dc81)
  • update discovery doc (8653ef7)

Bug Fixes

  • do not panic on a null element in a converted response slice or a null response body (c7c9c43), refs #899
  • overlays: fail CLI commands on failed interactions, redact tool headers, narrow ModelInteraction.required (34be134)
  • respect configured and default API version for Gemini API file uploads across SDKs (da10e09)
  • stop percent-encoding the Vertex resource prefix in Interactions URLs (4348fe0), refs #960

Miscellaneous Chores

1.77.0 (2026-10-29)

Features

  • Add support for continuation_token and requires_continuation status (51d761f)
  • add video upscale task parameter (cab8bf7)
  • interactions: add AutomaticContinuation to Interactions.Create in the Go SDK (0a6da26)
  • interactions: declare continuation_ids on Interaction (5b54979)
  • support video understanding in Gemini Developer API (5407437)

Bug Fixes

  • interactions: align SDK User-Agent and x-goog-api-client headers (58187ad)

1.76.0 (2026-10-22)

Features

  • Add store to GenerateContentConfig (Gemini Developer API only) and labels to LiveConnectConfig. (2ca47f8)
  • add base_voice to PromptedVoice in GAOS SDK (bfc9e1c)
  • Add generation_output_config to GenerateContentConfig (6ae1299)
  • Add Live API in-session context updates (send_context_update) and types across GenAI SDKs. (f1d0c6c)
  • include AudioContent, DocumentContent, and VideoContent in FunctionResultSubContent (68f1351)
  • interactions: populate output_text and output media helpers in the Go SDK (db2b842)
  • sdks: enable fixes.openSingleValueEnumsOct2026 across all SDKs (3892a92)
  • update discovery doc (525f275)

Bug Fixes

  • genai: raise mid-stream SSE errors and dropped connections consistently across SDKs; add continuation conformance tests (9aa2634)
  • interactions: sync InteractionSseEventInteractionStatus enum with InteractionStatus (7c129ed)
  • keep streamed chat turns in curated history when the stream ends with an empty chunk (4dd02c1), refs #795

1.75.0 (2026-10-14)

Features

  • interactions: add gemini-4-argon to Model enum (f8f7050)

Bug Fixes

  • clear continuation state and preserve prompt feedback when continuation hop returns no candidates (f0c1430)
  • fix the grounding metadata merging for automatic continuation in generate content and generate content stream (19689a2)
  • make continuation streams retry dropped hops, stop on prompt_feedback, and keep real errors in generate_content (0828ad4)

1.74.0 (2026-10-13)

Features

  • accumulate streaming continuation usage metadata and expose InitialPromptTokenCount (67c96f3)
  • Add continuation_token to the Interactions interaction.status_update event. (397c14f)
  • Add a CONTINUATION_REQUIRED Interaction status (125cd3e)
  • add billable token counts fields (43c5132)
  • Add startOffset/endOffset to Transcription and interimTranscriptTimestampEnabled to RealtimeInputConfig (3db3beb)
  • Add usage in InteractionStatusUpdate events (ba9bb94)
  • cli: update Speakeasy generator to v2.948.0, bump CLI to v0.4.2, and apply CLI overlay fixes (f660ee9)
  • Support automatic continuation (long decoding) (1f7179b)

Bug Fixes

  • clear intermediate CONTINUATION FinishReason in streaming continuation (2c08efa)
  • genai: deduplicate Flash and Omni descriptions in ModelOption (1957e9c)
  • Raise default HTTP timeouts so long-running requests aren't killed client-side (2e7f4aa)
  • resume streaming generation from the latest checkpoint continuation token on mid-stream errors or cutoffs (3e51164)
  • Update EnvironmentConfig.env schema to map keys to EnvVar or string. (b84d964)

1.73.0 (2026-10-07)

Features

  • add lyria-3.5, gemini-omni-1.1-flash, and gemini-omni-flash-preview to Interactions Model enum (0a17bc3)
  • go: expose Voices service on genai.Client and add Voices examples (770093f)
  • Support "allowlist": "disabled" in environment network config. (3a7595e)
  • support continuation_token in Interactions (5fed937)
  • update discovery doc (aeef05c)

Bug Fixes

  • gemini-api-cli: remove unsupported 2.5 and 1.6 models from Interactions model options and CLI examples (b1c58b2)

1.72.0 (2026-10-01)

Features

  • add gcs_metrics_uri for exporting tuning job metrics. (a2adbb5)
  • Add interactions (b85860b)
  • expose SpeechMetadata, VoiceConfig.voice, and SpeechAnnotation in public GenAI SDKs (c342626)
  • genai: refresh examples and add build tests for Go SDK examples (a8402d6)
  • implement environments files upload and download across .NET and Go SDKs (ee76a1c)
  • include labels for LiveClientSetup (69b3fa3)
  • support continuation_token in GenerateContent (3ecb7eb)
  • support labels for Gemini API GenerateContent (bf539c0)
  • update discovery doc (702e8d0)

Bug Fixes

  • interactions: ignore unchecked returns in mock server handlers in interaction_test.go (1617851)
  • Update GitHub Actions lint workflow for Node 24 compatibility (aeb441a)
  • use CreateAgentInteraction for Trigger.interaction and TriggerCreateParams.interaction (d918e85)

1.71.0 (2026-08-31)

Features

  • Add translation_config SDK support for GEAP. (83c445d)

1.70.0 (2026-08-25)

Features

  • Add AudioTranscriptionConfigMode (e601aac)

1.69.0 (2026-08-19)

Features

  • Add mode enum (VERBATIM, SMART) to AudioTranscriptionConfig and TranscriptionConfig. (5ca93e0)
  • Add enable_data_retention to ToolParallelAiSearch, Add step_count to ReinforcementTuningHyperParameters, Add BidiGenerateContentSetup (600a8cf)
  • Add IDLE state to live connection status enum and mark REQUIRES_ACTION as deprecated. (2bbd248)

1.68.0 (2026-08-12)

Features

  • Add interaction_status to LiveServerContent (b701130)

1.67.0 (2026-08-06)

Features

  • Add TOO_MANY_TOOL_CALLS to FinishReason enum. (e5a3783)
  • Support GoogleMaps Tool grounding_types places and routing (d76fedf)
  • Support HTTP retries in the Go GenAI SDK. (45c6782)

Bug Fixes

  • Add propertyOrdering auto-population for ResponseSchema and ResponseJsonSchema for Dotnet SDK (1e113a5)

1.66.0 (2026-07-28)

Features

  • [GenerateContentConfig] Add GenerationConfig.audio_transcription_config and Part.audio_transcription. (5d6dfc2)
  • Add flat language_codes field to AudioTranscriptionConfig. (3169529)
  • Allow api key + proj/location for enterprise mode (7357166)

1.65.0 (2026-07-21)

Features

  • A new field custom_vocabulary is added to message .google.cloud.aiplatform.v1beta1.BidiGenerateContentSetup (7d5b833)

Bug Fixes

  • add deprecation warnings to Imagen generate_images, edit_images, generate_videos (if using prompt/text/image args), LiveConnectConfig.GenerationConfig which will be removed in the next major version (d77df44)
  • Fix streaming truncation when HTTPOptions.Timeout is set. (3417576), closes #816

1.64.0 (2026-07-15)

Features

  • Add history config. (315c1b1)
  • Populate per-modality prompt token count in embedding responses for gemini-embedding-2 (7e186ba)

1.63.0 (2026-07-15)

Features

  • Add history config. (315c1b1)
  • Add response_format and Translation_config in GenerationConfig (caf2595)
  • Add Tool.exa_ai_search for Gemini Enterprise API (caf2595)
  • Populate per-modality prompt token count in embedding responses for gemini-embedding-2 (7e186ba)
  • voice consent signature types across all SDK languages. (ff4e260)

Bug Fixes

  • Fix exponential backoff bug in Go GenAI SDK. (55b57b9)

1.62.0 (2026-06-24)

Features

  • Add ComputerUse.disabled_safety_policies (c89e5eb)

1.61.0 (2026-06-17)

Features

  • Add audioOffset to VoiceActivity (43a352f)
  • Add interimInputTranscription to LiveServerContent (43a352f)
  • Add LanguageAuto, LanguageHints, and adaptationPhrases to AudioTranscriptionConfig (43a352f)

1.60.0 (2026-06-04)

Features

  • Add ServiceTier to UsageMetadata (74b1290)

1.59.0 (2026-06-03)

Features

  • Add Agent Platform MCP support to async generate_content (4b138c2)
  • Add transcription language code. (cc4dd9c)
  • Add TranslationConfig for live translation. (76f4126)
  • additional computer_use field support for vertex. (8831eb3)
  • Support 'additionalProperties', 'defs' and 'ref' in the GenerateContent.Schema type. (996b831)
  • Support Reinforcement Tuning in GenAI SDK (fecb49e)
  • Support ReinforcementTuning in GenAI SDK including ValidateReward API method. (c95d115)

1.58.0 (2026-05-21)

Features

  • add enable_prompt_injection_detection for Computer Use feature for the Gemini API. (19c2566)
  • add new fields (1608e80)

1.57.0 (2026-05-12)

Features

  • Expose Test and Server of the Replay API Client for the cross module usage in the AgentPlatform module. (27b142d)
  • support Blocking FunctionCall in Live API in AgentPlatform (Vertex) (e5e653c)
  • support speech to speech translation in Gemini Live (621b55d)

Bug Fixes

  • Fix version redaction logic in the replay API client. (af4c931)

1.56.0 (2026-05-04)

Features

  • [Python] Multimodal file search (c32ae6e)
  • Multimodal file search (989d064)

Bug Fixes

  • Simplify generated Go URL path formatting when VertexAI and MLDev URLs are identical. (d90a592), closes #761

1.55.0 (2026-04-29)

Features

  • Add output_info to BatchJob (9ec7d3b)
  • Add ImageResizeMode for GenerateVideos (8fc27f4)
  • Add InternalApplyConverterToSliceWithRoot to the common utilities. (7aaf1eb)
  • Add Vertex Dataset input and output options for batch jobs (73ee07c)
  • introduce BackendEnterprise and GOOGLE_GENAI_USE_ENTERPRISE env var (1181fe0)

1.54.0 (2026-04-13)

Features

  • Add "eu" as a supported service location for Vertex AI platform. (9245aba)
  • Add Live Avatar new fields (2ae252c)
  • Add webhook_config to batches.create() and models.generate_videos() (4790027)

1.53.0 (2026-04-08)

Miscellaneous Chores

1.52.1 (2026-03-31)

Bug Fixes

  • Fix service_tier enums. (5518442)

1.52.0 (2026-03-27)

Features

  • Add labels field to Veo configs (3fab63a)
  • Add model_status to GenerateContentResponse (Gemini API only) (8c040e3)
  • Add part_metadata in Part (Gemini API only) (8c040e3)
  • Add service tier to GenerateContent. (c07b6a0)
  • genai: add TURN_INCLUDES_AUDIO_ACTIVITY_AND_ALL_VIDEO to TurnCoverage (ea49f9a)
  • support hyperparameters in distillation tuning (b8873ea)
  • Support rendered_parts in GroundingSupport (8c040e3)

Bug Fixes

  • Fix empty InlinedEmbedContentResponses for inline batch embeddings in Go GenAI SDK (e70c9d1), closes #712
  • support us region routing (09d5be3)

1.51.0 (2026-03-17)

Features

  • Support include_server_side_tool_invocations for genai. (cec4bfd)

1.50.0 (2026-03-12)

Features

  • Add inference_generation_config to EvaluationConfig for Tuning (21b360f)
  • Add register_files (43a12bf)
  • Allow custom endpoints for authentication with Vertex AI in Go (0ebf016)
  • enable language code for audio transcription config in Live API for Vertex AI (59610d4)

1.49.0 (2026-02-28)

Features

  • Update data types from discovery doc. (37134ea)

1.48.0 (2026-02-25)

Features

  • Add Image Grounding support to GoogleSearch tool (ba34adf)
  • enable server side MCP and disable all other AFC when server side MCP is configured. (a550b3c)

1.47.0 (2026-02-18)

Features

  • Support multimodal embedding for Gemini Embedding 2.0 and support MaaS models in Models.embed_content() (Vertex AI API) (ef61dd1)

1.46.0 (2026-02-09)

Features

  • Support encryption_spec in tuning job creation configuration for GenAI SDK (025a3f0)

Bug Fixes

1.45.0 (2026-02-04)

Features

  • Update data types from discovery doc. (10f6ed9)

1.44.0 (2026-01-30)

Features

  • Support distillation tuning (cf4c39c)
  • Support OSS Tuning in GenAI SDK (9621775)

Bug Fixes

  • Add metadata in batch inlined response (f8e6b9b)

1.43.0 (2026-01-18)

Features

  • Add ModelArmorConfig support for prompt and response sanitization via the Model Armor service (580063f)
  • Update data types from discovery doc. (6842c63)
  • Update data types from discovery doc. (8065f85)

1.42.0 (2026-01-15)

Features

1.41.1 (2026-01-14)

Bug Fixes

  • Correct json tag typo in EntityLabel (096bff8)

1.41.0 (2026-01-07)

Features

  • [Python] add RegisterFiles so gcs files can be used with genai. (3062834)
  • Add gemini-3-pro-preview support for local tokenizer (1eeac81)
  • Add PersonGeneration to ImageConfig for Vertex Gempix (062e7e1)

1.40.0 (2025-12-16)

Features

  • Add minimal and medium thinking levels. (a7c0ed8)
  • Add ultra high resolution to the media resolution in Parts. (47b89eb)
  • ephemeral token support in go (7515bbe)
  • support multi speaker for Vertex AI (457541d)

1.39.0 (2025-12-10)

Features

  • Add enableEnhancedCivicAnswers feature in GenerateContentConfig (dd25df7)
  • Add IMAGE_RECITATION and IMAGE_OTHER enum values to FinishReason (33aa4f2)
  • Add voice activity detection signal. (326059c)

1.38.0 (2025-12-08)

Features

  • Support ReplicatedVoiceConfig (fd19393)

1.37.0 (2025-12-02)

Features

  • Add empty response for tunings.cancel() (ace417d)

1.36.0 (2025-11-17)

Features

  • add display name to FunctionResponseBlob (66bd0fb)
  • add display name to FunctionResponseFileData (f470cff)
  • Add generate_content_config.thinking_level (93b2586)
  • Add image output options to ImageConfig for Vertex (c4b28c3)
  • Add part.media_resolution (93b2586)
  • support Function call argument streaming for all languages (dd5ec01)

1.35.0 (2025-11-12)

Features

  • Add ToGenerationConfig method to GenerateContentConfig. fixes #506 (bd53df4)

Bug Fixes

  • Add missing fields to the model types (9e1b329)
  • Fix base_steps parameter for recontext_image (4f90d03)
  • Fix models.list() filter parameter (f5859fd)

1.34.0 (2025-11-05)

Features

  • Add UseDefaultCredentials method to ClientConfig. (d15baab)
  • add complete stats to BatchJob (0df87d7)
  • Add FileSearch tool and associated FileSearchStore management APIs (8ce95c8)
  • Add FileSearch tool and associated FileSearchStore management APIs (3e0a744)
  • Add image_size to ImageConfig (Early Access Program) (027f29c)
  • Added phish filtering feature. (eb849d1)

Bug Fixes

  • prevent nil pointer dereference panic in newAPIError function (251f7bf)
  • prevent nil pointer dereference panic in newAPIError function (#468) (ba15429)

1.33.0 (2025-10-29)

Features

  • Add safety_filter_level and person_generation for Imagen upscaling (3d7b57a)
  • Add support for preference optimization tuning in the SDK. (a6db7fe)
  • Pass file name to the backend when uploading with a file path (6b5e4f4)
  • support default global location when not using api key with vertexai backend (44f01d6)

1.32.0 (2025-10-21)

Features

  • Add enable_enhanced_civic_answers in GenerationConfig (c876512)
  • support jailbreak in HarmCategory and BlockedReason (65e1efc)

Bug Fixes

  • change back SingleEmbedContentResponse.TokenCount to int64 due to mistake in #557 (e05f15d)
  • do not append project and client for Vertex AI when using API key (c27f13a), closes #564

1.31.0 (2025-10-15)

Features

  • support CreateEmebddings in batches.go (a488476)
  • Support video extension for Veo on Gemini Developer API (14ecba9)

1.30.0 (2025-10-10)

Features

  • Enable Google Maps tool for Genai. (9aba5c7)
  • Support enableWidget feature in GoogleMaps (cd1e6b7)
  • Support Gemini batch inline request's metadata and add test coverage to safety setting (f12f353)

1.29.0 (2025-10-08)

Features

  • Add labels field to Imagen configs (d01fe5e)
  • Add utility methods for creating FunctionResponsePart and creating FunctionResponse Part with FunctionResponseParts (10de2ab)
  • Enable Ingredients to Video and Advanced Controls for Veo on Gemini Developer API (Early Access Program) (3165554)

1.28.0 (2025-10-02)

Features

  • Add NO_IMAGE enum value to FinishReason (4f65f45)
  • Add thinking_config for live (54152b2)

1.27.0 (2025-10-01)

Features

  • Add ImageConfig to GenerateContentConfig (55399fd)
  • rename ComputerUse tool (early access) (d976966)

Bug Fixes

  • fix system_instruction mapping issue in batches module (c862a6d)

1.26.0 (2025-09-25)

Features

  • Add FunctionResponsePart & ToolComputerUse.excludedPredefinedFunctions (8b97cef)
  • Support Imagen 4 Ingredients on Vertex (a9ab364)

Bug Fixes

  • Expose JOB_STATE_RUNNING and JOB_STATE_EXPIRED for Gemini Batches states (26e0182)
  • fix systemInstruction mapping issue in batch generate content. fixes #505 (3997ea2)
  • initialization of pre_tuned_model_checkpoint_id from tuning config. (b093bcf)
  • Prevent adding ,string JSON tag for int64/uint64 fields in Schema class. fixes #511 #481 (3423dd3)

1.25.0 (2025-09-16)

Features

  • Add 'turn_complete_reason' and 'waiting_for_input' fields. (2d48288)
  • Add VideoGenerationMaskMode enum for Veo 2 Editing (65f9058)
  • local tokenizer for text (ee46a52)

1.24.0 (2025-09-09)

Features

  • [Python] Implement async embedding batches for MLDev. (f32fb26)
  • Add labels to create tuning job config (c13a2a5)
  • generate the function_call class's converters (995a3ac)
  • Support Veo 2 Editing on Vertex (7fd6940)

Bug Fixes

  • Enable id field in FunctionCall for Vertex AI. (a3f3c2b)

1.23.0 (2025-09-02)

Features

  • Add resolution field for Gemini Developer API Veo 3 generation (b6a989c)

1.22.0 (2025-08-27)

Features

  • add sdkHttpResponse.headers to *Delete responses. (ac0e763)
  • Add add_watermark field for recontext_image (Virtual Try-On, Product Recontext) (07b6f57)
  • Add output_gcs_uri to Imagen upscale_image (c08d9f3)
  • Add VALIDATED mode into FunctionCallingConfigMode (c282e79)
  • Add VideoGenerationReferenceType enum for generate_videos (635b825)
  • refactor Go SDK to use pointers for optional parameters (3ff328a)
  • support tunings.cancel in the genai SDK for Python, Java, JS, and Go (8c46fd2)

1.21.0 (2025-08-18)

Features

  • Support Imagen image segmentation on Vertex (2a38843)
  • Support Veo 2 Reference Images to Video Generation on Vertex (9894324)

Bug Fixes

  • Add a missing resp.Body.Close() to deserializeStreamResponse. (bf3fb3f)

1.20.0 (2025-08-13)

Features

  • enable continuous fine-tuning on a pre-tuned model in the SDK. (1f20493)
  • support document name in grounding metadata (f673e20)
  • Support exclude_domains in Google Search and Enterprise Web Search (2547ad1)

1.19.0 (2025-08-06)

Features

  • Add image_size field for Gemini Developer API Imagen 4 generation (3ccd2b0)
  • Add parts length check (#444) (a4896f3)
  • allow methods in batch to return headers in sdk_http_response by default (3a1d6d8)
  • enable responseId for Gemini Developer API (9845a91)
  • support curated history in the Go chats module. (2b99b83)
  • Support image recontext on Vertex (fc7ee78)
  • Support new enum types for UrlRetrievalStatus (ef63a73)

1.18.0 (2025-07-30)

Features

  • support response headers in Go for all methods. (4865b93)

1.17.0 (2025-07-17)

Features

  • Add generateVideosFromSource in Go, refactor private generateVideos (cd625ec)
  • Add image_size field for Vertex Imagen 4 generation (5458a1c)
  • Support HTTPOptions.ExtraBody (bce0d4a)

1.16.0 (2025-07-16)

Features

  • Add addWatermark parameter to the edit image configuration. (d7e2847)
  • add Tuning support for Go (0c90d6c)
  • Migrate Go file.create method to use the sdk_http_response field and remove http_headers (082b7c7)

Bug Fixes

  • live: Enhance security by moving api key from query parameters to header (7cc7b7c)

1.15.0 (2025-07-09)

Features

  • Add new languages for Imagen 4 prompt language (afeabc2)
  • make credentials optional when providing HTTPClient (8b63004)

1.14.0 (2025-07-01)

Features

  • Enable HttpOptions timeout (1ca3aaf)
  • Support Batches delete (b2cf7bb)
  • Support different media input in Vertex Live API (b6650b5)

1.13.0 (2025-06-25)

Features

  • Add compressionQuality enum for generate_videos (061e567)
  • Add enhance_input_image and image_preservation_factor fields for upscale_image (6b9e07f)
  • Batches support in Go (dcae33f)
  • expose the responseJsonSchema in GenerateContentConfig (611adde)

1.12.0 (2025-06-18)

Features

  • enable json schema for controlled output and function declaration. (f2abe6b)

1.11.1 (2025-06-13)

Bug Fixes

  • Exclude video field from GenerateVideos to restore backwards compatibility in Go SDK (a868b8b)
  • Fix precedence when reading from environment variables api key and project for Go SDK (46bf8cc)

1.11.0 (2025-06-11)

Features

  • Add datastore_spec field for VertexAISearch (489ee80)
  • Add support for Veo frame interpolation and video extension (eae965b)
  • RAG - Introducing context storing for Gemini Live API. (b752aa0)
  • Support API keys for VertexAI mode for Go SDK (3b4aadf)

Bug Fixes

  • handle structured error in the stream chunk. fixes #355 (fb361ac)

1.10.0 (2025-06-05)

Features

  • add extras request provider to the HTTP options (7a00367)

Bug Fixes

  • Merge ExtrasRequestProvider for client level config and function level config (f63de00)

1.9.0 (2025-06-04)

Features

  • Add enhance_prompt field for Gemini Developer API generate_videos (04c9207)
  • Enable url_context for Vertex (6ca5b9c)
  • go: Support GEMINI_API_KEY as environment variable for setting API key. (83ba1e0)

1.8.0 (2025-05-30)

Features

  • Adding thought_signature field to the Part to store the signature for thoughts. (080fd90)
  • include UNEXPECTED_TOOL_CALL enum value to FinishReason for Vertex AI APIs. (03f0ea1)

Bug Fixes

  • Rename LiveEphemeralParameters to LiveConnectConstraints. (e7c5ee7)

1.7.0 (2025-05-28)

Features

  • Add generate_audio field for private testing of video generation (d48d6f3)
  • support new fields in FileData, GenerationConfig, GroundingChunkRetrievedContext, RetrievalConfig, Schema, TuningJob, VertexAISearch, (9331c82)

Bug Fixes

  • use correct mimetype for image content (#301) (ddc69b8)

1.6.0 (2025-05-19)

Features

  • add time range filter to Google Search Tool (02bec9d)
  • Add basic support for async function calling. (514cf37)
  • add live proactivity_audio and enable_affective_dialog (a72f5ce)
  • add multi-speaker voice config (aae87a9)
  • Add support for lat/long in search. (0e2ba95)
  • Add Video FPS, and enable start/end_offset for MLDev (fa403ac)
  • support customer-managed encryption key in cached content (a8a6dc2)
  • Support Url Context Retrieval tool (7bf9acc)

Bug Fixes

  • fix SendMessageStream when iterator callback returns false. fixes #310 (08cf7d9)

1.5.0 (2025-05-13)

Features

  • Add Send and SendStream to chats module. Related to #295 (f2f8041)
  • support display_name for Blob class when calling Vertex AI (10b5438)
  • Support tuning checkpoints (c65e033)

1.4.0 (2025-05-08)

Features

  • Add Tool.enterprise_web_search field (452d379)
  • Add support for Grounding with Google Maps (76c6472)
  • enable input transcription for Gemini API. (f1ccf67)

Bug Fixes

  • add retry logic for missing X-Goog-Upload-Status header for golang (1a25f15)

1.3.0 (2025-04-30)

Features

  • add models.delete and models.update to manage tuned models (6f5bbed)
  • add NewPartFromFile for File type convenience. Related to #281 (2ac0429)
  • add support for live grounding metadata (9ce2ed9)
  • make min_property, max_property, min_length, max_length, example, patter fields available for Schema class when calling Gemini API (b487724)
  • Populate X-Server-Timeout header when a request timeout is set. (8f446a0)
  • Support setting the default base URL in clients via SetDefaultBaseURLs() (f465a20)

Bug Fixes

  • do not raise error for default field in Schema for Gemini API calls (ec31e4b)
  • do not remove content parts with Text unset. (b967057)
  • files: deep copy config struct before modifying it. (a6b0fd6)

1.2.0 (2025-04-23)

Features

  • add additional realtime input fields (1190539)
  • Expose transcription configurations for audio in TS, and move generationConfig to the top level LiveConnectConfig (ead7e49)
  • support default field in Schema when users call Gemini API (643eb80)

1.1.0 (2025-04-16)

Features

  • Add converters to support continuous sessions with a sliding window (97bbba4)
  • add support for model_selection_config to GenerateContentConfig (a44a6c7)
  • Add types for configurable speech detection (658c17a)
  • Support audio transcription in Vertex Live API (f09dfab)
  • Support RealtimeInputConfig, and language_code in SpeechConfig in python (f90a0ec)
  • Update VertexRagStore (62d582c)

Bug Fixes

  • files: use io.ReadFull to read the file. fixes #237 (908c783)
  • Fix error "assignment to entry in nil map" of Files.Upload() when config.HTTPOptions is nil (#235) (05c0c49)
  • fix MIME type error in UploadFromPath and add unit tests. fixes: #247 (f851639)

1.0.0 (2025-04-09)

⚠ BREAKING CHANGES

  • Support SendClientContent/SendRealtimeInput/SendToolResponse methods in Session struct and remove Send method
  • Merge GenerationConfig to LiveConnectConfig. GenerationConfig is removed.
  • Change NewContentFrom... functions role param type from string to Role and miscs docstring improvements
  • Change some pointer to value type and value to pointer type

Features

  • Add domain to Web GroundingChunk (183ac49)
  • Add generationComplete notification to Live ServerContent (9a038b9)
  • Add session resumption to Live module (4a92461)
  • add session resumption. (507137b)
  • Add support for Chats streaming in Go SDK (9ee0523)
  • Add thinking_budget to ThinkingConfig for Gemini Thinking Models (f811ee4)
  • Add traffic type to GenerateContentResponseUsageMetadata (601add2)
  • Add types for configurable speech detection (f4e1b11)
  • Add types to support continuous sessions with a sliding window (5d4f5d7)
  • Add UsageMetadata to LiveServerMessage (4286c6b)
  • expose generation_complete, input/output_transcription & input/output_audio_transcription to SDK for Vertex Live API (0dbbc82)
  • Merge GenerationConfig to LiveConnectConfig. GenerationConfig is removed. (65b7c1c)
  • Remove experimental warnings for generate_videos and operations (2e4bb0b)
  • Support files delete, get, list, download/ (8e7b3fd)
  • Support files upload method (ce790dd)
  • support media resolution (825c81d)
  • Support SendClientContent/SendRealtimeInput/SendToolResponse methods in Session struct and remove Send method (c8ecaf4)
  • use io.Reader in Upload function and add a new convenience function UploadFromPath. fixes #222 (1c064e3)

Bug Fixes

  • Change NewContentFrom... functions role param type from string to Role and miscs docstring improvements (7810e07)
  • Change some pointer to value type and value to pointer type (0d2ba97)
  • fix Add() dead loop (afa2324)
  • Fix failing chat_test (aebbdaa)

0.7.0 (2025-03-31)

⚠ BREAKING CHANGES

  • Add error return type to Close() function
  • consolidate NewUserContentFrom* and NewModelContentFrom* functions into NewContentFrom* to make API simpler
  • Support quota project and migrate ClientConfig.Credential from google.Credentials to auth.Credential type.
  • Change caches TTL field to duration type.
  • rename ClientError and ServerError to APIError. fixes: #159

Features

  • Add Chats module for Go SDK (non-stream only) (e7f75fd)
  • Add engine to VertexAISearch (cc2ab5d)
  • add IMAGE_SAFTY enum value to FinishReason (cc6081a)
  • add MediaModalities for ModalityTokenCount (0969afd)
  • Add Veo 2 generate_videos support in Go SDK (5321a25)
  • allow title property to be sent to Gemini API. Gemini API now supports the title property, so it's ok to pass this onto both Vertex and Gemini API. (8f27aba)
  • consolidate NewUserContentFrom* and NewModelContentFrom* functions into NewContentFrom* to make API simpler (e8608b1)
  • merge GenerationConfig into LiveConnectConfig (96232de)
  • rename ClientError and ServerError to APIError. fixes: #159 (12adbfa)
  • Save prompt safety attributes in dedicated field for generate_images (eb3cfdc)
  • support new UsageMetadata fields (3a56c63)
  • Support quota project and migrate ClientConfig.Credential from google.Credentials to auth.Credential type. (74c05fb)

Bug Fixes

  • Add error return type to Close() function (673a7f7)
  • Change caches TTL field to duration type. (11271b4)
  • fix list models API url (036c4d3)
  • fix response modality in streaming mode. fixes #163. fixes #158 (996dac3)
  • missing zero value bug in setValueByPath. fixes #196 (557c6d8)
  • schema transformer logic fix. (8017092)
  • use snake_case in embed_content request/response parsing. fixes #174 (ba644e1)

Miscellaneous Chores

0.6.0 (2025-03-19)

⚠ BREAKING CHANGES

  • support duration type and remove NewPartFromVideoMetadata function
  • Change *time.Time type to time.Time.
  • remove error from the GenerateContentResponse.Text() return values and add more samples(text embedding, tokens, models)
  • change GenerateImageConfig.NumberOfImages to value type. And add clearer error message and docstring to other APIs.
  • Remove default role to "user" for GenerateContent and GenerateContentStream.

Features

  • Add base steps to EditImageConfig (e3c8252)
  • Change *time.Time type to time.Time. (d554a08)
  • change GenerateImageConfig.NumberOfImages to value type. And add clearer error message and docstring to other APIs. (a75a9ae)
  • enable union type for Schema when calling Gemini API. (2edcc55)
  • Remove default role to "user" for GenerateContent and GenerateContentStream. (74d4647)
  • remove error from the GenerateContentResponse.Text() return values and add more samples(text embedding, tokens, models) (1dc5c1c)
  • support duration type and remove NewPartFromVideoMetadata function (b2a13ab)
  • Support global endpoint in go natively (a29b806)
  • Support returned safety attributes for generate_images (cc2bf1a)

Bug Fixes

  • Make month and day optional for PublicationDate. fixes #141 (8a61516)
  • Remove unsupported parameter negative_prompt from Gemini API generate_images (be2619d)

Miscellaneous Chores

0.5.0 (2025-03-06)

⚠ BREAKING CHANGES

  • change int64, float64 types to int32, unit32, float32 to prevent data loss
  • remove ClientConfig.Timeout and add HTTPOptions to ...Config structs

Features

  • Add Headers field into HTTPOption struct (5ec9ff4)
  • Add response_id and create_time to GenerateContentResponse (f46d996)
  • added Models.list() function (6c2eae4)
  • enable minItem, maxItem, nullable for Schema type when calling Gemini API. (fb6c8a5)
  • enable quick accessor of executable code and code execution result in GenerateContentResponse (21ca251)
  • remove ClientConfig.Timeout and add HTTPOptions to ...Config structs (ba6c431)
  • Support aspect ratio for edit_image (06d554f)
  • support edit image and add sample for imagen (f332cf2)
  • Support Models.EmbedContent function (a71f0a7)

Bug Fixes

  • change int64, float64 types to int32, unit32, float32 to prevent data loss (af83fa7)
  • log warning instead of throwing error for GenerateContentResponse.text() quick accessor when there are mixed types of parts. (006e3af)

Miscellaneous Chores

0.4.0 (2025-02-24)

Features

  • Add Imagen upscale_image support for Go (8e2afe9)
  • introduce usability functions to allow quick creation of user content and model content. (12b5dee)
  • support list all caches in List and All functions (addc388)
  • support Models .Get, .Update, .Delete (e67cd8b)

Bug Fixes

  • fix the civil.Date parsing in Citation struct. fixes #106 (f530fcf)
  • missing context in request. fixes #104 (747c5ef)
  • Remove request body when it's empty. (cfc82e3)

0.3.0 (2025-02-12)

Features

  • Enable Media resolution for Gemini API. (a22788b)
  • support property_ordering in response_schema (fixes #236) (ac45038)

0.2.0 (2025-02-05)

Features

  • Add enhanced_prompt to GeneratedImage class (449f0fb)
  • Add labels for GenerateContent requests (98231e5)

Bug Fixes

  • remove unsupported parameter from Gemini API (39c8868)
  • Use camel case for Go function parameters (94765e6)

0.1.0 (2025-01-29)

⚠ BREAKING CHANGES

  • Make some numeric fields to pointer type and bool fields to value type, and rename ControlReferenceTypeControlType* constants

Features

  • [genai-modules][models] Add HttpOptions to all method configs for models. (765c9b7)
  • Add Imagen generate_image support for Go SDK (068fe54)
  • Add support for audio_timestamp to types.GenerateContentConfig (fixes #132) (cfede62)
  • Add support for enhance_prompt to model.generate_image (a35f52a)
  • Add ThinkingConfig to generate content config. (ad73778)
  • enable Text() and FunctionCalls() quick accessor for GenerateContentResponse (3f3a450)
  • Images - Added Image.mime_type (3333511)
  • introducing HTTPOptions to Client (e3d1d8e)
  • make Part, FunctionDeclaration, Image, and GenerateContentResponse classmethods argument keyword only (f7d1043)
  • Make some numeric fields to pointer type and bool fields to value type, and rename ControlReferenceTypeControlType* constants (ee4e5a4)
  • support caches create/update/get/update in Go SDK (0620d97)
  • support usability constructor functions for Part struct (831b879)

Miscellaneous Chores

Code Refactoring

  • rename GenerateImage() to GenerateImage(), rename GenerateImageConfig to GenerateImagesConfig, rename GenerateImageResponse to GenerateImagesResponse, rename GenerateImageParameters to GenerateImagesParameters (ebb231f)

0.0.1 (2025-01-10)

Features

  • enable response_logprobs and logprobs for Google AI (#17) (51f2744)
  • Go SDK Live module implementation for GoogleAI backend (f88e65a)
  • Go SDK Live module initial implementation for VertexAI. (4d82dc0)

Bug Fixes

  • change string type to numeric types (bfdc94f)
  • fix README typo (5ae8aa6)