1.2.9

feat: support output dimensions for embedding models (#490) (#522)

Add optional outputDimension field for embedding models to request
specific dimensions from providers that support Matryoshka
Representation Learning (MRL).

Supported providers:
- OpenAI text-embedding-3-* (via 'dimensions' parameter)
- Google gemini-embedding-001 (via 'config.outputDimensionality')

Changes:
- Add outputDimension field to EmbeddingModel type
- Update all provider getEmbedding methods to accept dimensions option
- Add Output Dimensions input field in Add Embedding Model modal
- Validate that returned dimension matches requested outputDimension

1.2.8

Add Gemini Plan (OAuth) support (#513)

* feat: Add Gemini plan provider with OAuth support

- Introduced Gemini plan as a new provider type with OAuth integration.
- Implemented authentication flow including token exchange and callback server.
- Updated settings schema to include Gemini provider and associated OAuth fields.
- Enhanced chat model settings to support Gemini-specific configurations.
- Added migration logic to transition existing settings to support the new provider.
- Comprehensive tests added for migration and Gemini functionality.

* chore: Update Gemini plan modal and connection descriptions

* docs: Update README and settings to warn about risks of third-party OAuth access for Claude subscriptions

- Added a warning in the README about the risks associated with connecting a Claude subscription, including potential account bans.
- Introduced a warning message in the PlanConnectionsSection to inform users about the restrictions on third-party OAuth access and advised caution in usage.
- Enhanced styles for warning messages to improve visibility and user awareness.

* fix: Fix lint

* fix: Resolve coderabbitai review

1.2.7

Add plan-based Claude/OpenAI connections with OAuth (#509)

* feat: Add OpenAI Codex provider and OAuth integration

- Introduced OpenAI Codex as a new provider type with OAuth support.
- Implemented authentication flow including token exchange and callback server.
- Updated settings schema to include Codex provider and associated OAuth fields.
- Enhanced chat and provider management to accommodate Codex integration.
- Added migration logic to transition existing settings to support the new provider.
- Comprehensive tests added for migration and Codex functionality.

* feat: Enhance HTTP transport for Codex integration

- Added comments explaining the use of Node's http/https modules for Codex endpoints due to CORS restrictions.
- Updated the nodePost function to throw an error when called on mobile platforms, ensuring proper platform handling.
- Refactored the import of IncomingMessage type for better clarity and consistency.

* feat: Add Anthropic Claude Code provider with OAuth support

- Introduced the Anthropic Claude Code provider, including OAuth integration for authentication.
- Implemented the necessary authentication flow, including token exchange and authorization URL generation.
- Updated settings schema to include the new provider type and associated OAuth fields.
- Enhanced provider management and chat model handling to accommodate Claude Code integration.
- Added migration logic to transition existing settings to support the new provider.
- Comprehensive tests added for migration and Claude Code functionality.

* chore: Update OpenAI package and refactor message adapters

- Upgraded OpenAI package from version 4.91.1 to 6.8.1 in package.json and package-lock.json.
- Refactored DeepSeekMessageAdapter and PerplexityMessageAdapter to normalize tool calls using the new method.
- Enhanced OpenAIMessageAdapter with a new normalizeToolCalls method to streamline tool call handling.
- Updated NoStainlessOpenAI to improve header management in requests.

* feat: Add OpenAI Codex model settings and reasoning support

- Introduced settings for the OpenAI Codex model, allowing users to configure reasoning effort and summary.
- Enhanced CodexMessageAdapter to handle reasoning summary extraction and inclusion in responses.
- Updated OpenAICodexProvider to normalize requests with reasoning parameters.
- Modified chat model schema to accommodate new reasoning fields.
- Added utility functions for reasoning summary extraction in response payloads.

* feat: Update ChatModelSettings to support new Anthropic Claude Code model

* feat: Refactor provider settings and enhance migration logic for OpenAI and Anthropic models

- Reintroduced OpenAI provider settings in the constants file for consistency.
- Updated DEFAULT_PROVIDERS and DEFAULT_CHAT_MODELS to include new models for OpenAI Codex and Anthropic Claude Code.
- Enhanced migration logic to ensure proper transition to version 15, incorporating new default providers and chat models.

* feat: Refactor provider types and update settings for OpenAI and Anthropic models

- Replaced 'openai-codex' and 'anthropic-claude-code' with 'openai-plan' and 'anthropic-plan' in constants and provider settings.
- Updated DEFAULT_PROVIDERS and DEFAULT_CHAT_MODELS to reflect new provider types.
- Adjusted migration logic to ensure compatibility with the new provider types.
- Enhanced related components and schemas to support the updated provider structure.

* feat: Implement subscription connection settings for OpenAI and Anthropic plans

- Added new PlanConnectionsSection to manage connections for OpenAI and Claude subscriptions.
- Introduced ConnectClaudePlanModal and ConnectOpenAIPlanModal for OAuth authentication flows.
- Enhanced styles for plan connection cards and status indicators.
- Updated SettingsTabRoot to include the new PlanConnectionsSection.
- Refactored ProvidersSection to exclude subscription providers from API key management.

* feat: Add PLAN_PROVIDER_TYPES constant and update provider filtering logic

* refactor: Clean up provider connection logic and improve formatting

* chore: Update .gitignore to exclude CLAUDE.md

* style: Enhance layout and descriptions in PlanConnectionsSection

* fix: Resolve coderabbit review

- Refactored URL parameter extraction to support dynamic keys for better reusability.
- Added state validation to ensure OAuth state consistency during the connection process.
- Updated error messages for clearer user guidance in case of missing or mismatched OAuth states.
- Introduced a new utility function for posting form data to streamline token requests.
- Enhanced HTTP transport methods to support form URL encoding and improved error handling.

* chore: Update column header in ProvidersSection from 'Credential' to 'API Key'

* chore: Update README for v1.2.7 release

* test: Enhance migration tests for v14 to v15

1.2.6

fix: resolve npm audit vulnerabilities (#507)

* fix: resolve npm audit vulnerabilities

1.2.5

Add new models: sonnet-4.5 and haiku-4.5 (#492)

* feat: Update chat models and implement migration from v11 to v12

- Updated default chat model ID to 'claude-sonnet-4.5' and added new recommended models.
- Introduced migration logic to transition settings from version 11 to 12, merging existing chat models with new defaults and adding 'claude-sonnet-4.5' and 'claude-haiku-4.5'.
- Removed outdated models: 'claude-sonnet-4.0', 'claude-3.7-sonnet', 'claude-3.5-sonnet', and 'claude-3.5-haiku'.
- Added tests to verify migration functionality.

* docs: Update README.md with maintenance notice and contribution invitation

- Added a maintenance notice indicating the plugin is not under active development and may have delayed responses to issues and feature requests.
- Included an invitation for community contributions and collaboration.

1.2.4

Update README.md

1.2.3

fix: remove tools field in the DeepSeek HTTP request. (#398)

* fix: remove tools field in the DeepSeek HTTP request.

* polish: modify by lint

* feat: Enhance DeepSeekMessageAdapter to include tool_calls in message handling

- Added tool_calls property to both message and delta processing in DeepSeekMessageAdapter for improved tool integration.
- Updated responseGenerator to handle cases where no tools are available, setting tools to undefined to prevent errors.

---------

Co-authored-by: Kevin On <40454531+kevin-on@users.noreply.github.com>

1.2.2

Update models: GPT-5, Claude 4, Gemini 2.5 (#452)

* feat: Update anthropic models and add thinking toggle in model settings

* feat: refactor OpenAI reasoning feature with improved UX and GPT-5 support

- Replace simple reasoning_effort field with structured reasoning object
- Add reasoning.enabled toggle for better user control
- Support GPT-5, GPT-5-mini, and GPT-5-nano models with pricing

* feat: Update Gemini models

- Update Gemini models: remove 1.5 versions, add 2.5-pro and 2.5-flash-lite
- Remove Groq pricing constants and cost calculation logic

* feat: Add model fallback

- Add automatic fallback to first available chat model when selected model is not found
- Create LLMModelNotFoundException for better error handling

* feat: Add migration from v10 to v11 with new model structure

- Add migration from v10 to v11 with new reasoning/thinking structure
- Transform OpenAI models with reasoning_effort to new reasoning object structure
- Transform Anthropic models with thinking.budget_tokens to new thinking structure
- Add new default models: claude-sonnet-4.0, claude-opus-4.1, gpt-5 series, gemini-2.5 series

* refactor: Centralize default model ids and update schema fallbacks/recommendations

- Add DEFAULT_CHAT_MODEL_ID ('claude-sonnet-4.0') and DEFAULT_APPLY_MODEL_ID ('gpt-4.1-mini') in src/constants.ts
- Update RECOMMENDED_MODELS_FOR_CHAT to ['claude-sonnet-4.0', 'gpt-4.1'] and RECOMMENDED_MODELS_FOR_APPLY to ['gpt-4.1-mini']
- Update smartComposerSettingsSchema to use the new default IDs as fallbacks (src/settings/schema/setting.types.ts)
- Adjust tests to assert defaults via constants (src/settings/schema/settings.test.ts)

* fix: Resolve CodeRabbit review

1.2.1

feat: Add Mistral provider support (#376)

1.2.0

refactor: Update styles and improve component structure (#373)

* refactor: Update styles and improve component structure

- Adjusted CSS styles for better layout and spacing in chat components.
- Refactored ChatUserInput component to streamline rendering of mentionable badges.
- Modified MCPManager to ensure subscribers are notified after cache invalidation.
- Updated request type definitions to enforce properties in FunctionDescription.
- Enhanced response generation to ensure parameters are correctly structured.

* feat: Add settings callout for tool usage warnings

- Introduced a new callout in the MCP settings section to inform users about potential costs associated with tool responses.
- Updated CSS styles for the callout to enhance visibility and layout.