* fix: error when updating bookmarks if no query data
* feat: localize bookmark dialog, form labels and validation messages, also improve validation
* feat: add localization for EmptyPromptPreview component and update translation.json
* chore: add missing localizations for static UI text
* chore: update AgentPanelContextType and useGetAgentsConfig to support null configurations
* refactor: update agent categories to support localization and custom properties, improve related typing
* ci: add localization for 'All' category and update tab names in accessibility tests
* chore: remove unused AgentCategoryDisplay component and its tests
* chore: add localization handling for agent category selector
* chore: enhance AgentCard to support localized category labels and add related tests
* chore: enhance i18n unused keys detection to include additional source directories and improve handling for agent category keys
* refactor: Resource Migration Scripts for DocumentDB compatibility
* fix: Correct type annotation for `db` parameter in ensureCollectionExists function
* fix(latex): prevent LaTeX conversion when closing $ is preceded by backtick
When text contained patterns like "$lookup namespace" followed by "`$lookup`",
the regex would match from the first $ to the backtick's $, treating the entire
span as a LaTeX expression. This caused programming constructs to be incorrectly
converted to double dollars.
- Added negative lookbehind (?<!`) to single dollar regex
- Prevents matching when closing $ immediately follows a backtick
- Fixes issues with inline code blocks containing $ symbols
* fix(latex): detect currency amounts with 4+ digits without commas
The currency regex pattern \d{1,3} only matched amounts with 1-3 initial digits,
causing amounts like $1157.90 to be interpreted as LaTeX instead of currency.
This resulted in text like "$1157.90 (text) + $500 (text) = $1657.90" being
incorrectly converted to a single LaTeX expression.
- Changed pattern from \d{1,3} to \d+ to match any number of initial digits
- Now properly escapes $1000, $10000, $123456, etc. without requiring commas
- Maintains support for comma-formatted amounts like $1,234.56
* fix(latex): support currency with unlimited decimal places
The currency regex limited decimal places to 1-2 digits (\.\d{1,2}), which
failed to properly escape amounts with more precision like cryptocurrency
values ($0.00001234), gas prices ($3.999), or exchange rates ($1.23456).
- Changed decimal pattern from \.\d{1,2} to \.\d+
- Now supports any number of decimal places
- Handles edge cases like scientific calculations and high-precision values
* fix: available tools retrieval with correct mcp caching and conversion
* test: Enhance PluginController tests with MCP tool mocking and conversion
* refactor: Simplify PluginController tests by removing unused mocks and enhancing test clarity
* feat: Add directEndpoint option to OpenAIConfigOptions and update fetch logic to override /chat/completions URL
* feat: Add directEndpoint support to fetchModels and update loadConfigModels logic
* chore: import paths for isEnabled and logger in title.js
* ⛔ fix: `AbortSignal` Cleanup Logic for New Chats
* test: Add `isNewConvo` parameter to onStart expectation in BaseClient tests
Fixes "Cannot read properties of undefined (reading 'key')" error in parameter panels by filtering out null/undefined values before mapping operations.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-authored-by: Claude <noreply@anthropic.com>
* decouple MCP dialog from BadgeRowContext
* chore: import order and style according to guidelines
---------
Co-authored-by: Dustin Healy <dustinhealy1@gmail.com>
* 📦 chore: Bump `@librechat/api` to v1.3.3
* 📦 chore: Bump `librechat-data-provider` to v0.8.003
* 📦 chore: Bump `@librechat/data-schemas` to v0.0.18
* 📦 chore: Bump `@librechat/client` to v0.2.6
* 📦 chore: Update workflow name for `librechat-data-provider` and add manual dispatch
* 📦 chore: Update Node.js version to 20 in data provider workflow
* feat: Add conversation ID support to custom endpoint headers
- Add LIBRECHAT_CONVERSATION_ID to customUserVars when provided
- Pass conversation ID to header resolution for dynamic headers
- Add comprehensive test coverage
Enables custom endpoints to access conversation context using {{LIBRECHAT_CONVERSATION_ID}} placeholder.
* fix: filter out unresolved placeholders from headers (thanks @MrunmayS)
* feat: add support for request body placeholders in custom endpoint headers
- Add {{LIBRECHAT_BODY_*}} placeholders for conversationId, parentMessageId, messageId
- Update tests to reflect new body placeholder functionality
* refactor resolveHeaders
* style: minor styling cleanup
* fix: type error in unit test
* feat: add body to other endpoints
* feat: add body for mcp tool calls
* chore: remove changes that unnecessarily increase scope after clarification of requirements
* refactor: move http.ts to packages/api and have RequestBody intersect with Express request body
* refactor: processMCPEnv now uses single object argument pattern
* refactor: update processMCPEnv to use 'options' parameter and align types across MCP connection classes
* feat: enhance MCP connection handling with dynamic request headers to pass request body fields
---------
Co-authored-by: Gopal Sharma <gopalsharma@gopal.sharma1>
Co-authored-by: s10gopal <36487439+s10gopal@users.noreply.github.com>
Co-authored-by: Dustin Healy <dustinhealy1@gmail.com>
* fix: register openId the strategy if setupOpenId succeeded
* chore: linting and update imports
* refactor: extract OpenID configuration into a separate function
---------
Co-authored-by: Denis <denis.sheremetov@gmail.com>
* ✨ feat: Add support for enabling Redis cluster configuration
* ✨ feat: Enhance Redis client initialization to support cluster configuration without multiple URIs
* ✨ feat: Add tests for USE_REDIS_CLUSTER configuration and validation
* 🐞 fix: Remove unnecessary blank line in cacheConfig tests
* adding beta header context-1m-2025-08-07 to claude sonnet 4 to increase contact window to 1M tokens
* adding context-1m beta header to test cases
* ci: Update Anthropic `getLLMConfig` tests and headers for model variations
- Refactored test cases to ensure proper handling of model variations for 'claude-sonnet-4'.
- Cleaned up unused mock implementations in tests for better clarity and performance.
* refactor: regex in header retrieval for 'claude-sonnet-4' models
* refactor: default tokens for 'claude-sonnet-4' to `1,000,000`
* refactor: add token value retrieval and pattern matching to model tests
---------
Co-authored-by: Dirk Petersen <no-reply@nowhere.com>
* refactor: `packages/api` build scripts for better inline debugging
* refactor: Explicitly select secure fields as no longer returned by default, exclude backupCodes from user data retrieval in authentication and 2FA processes
* refactor: Backup Codes UI to not expect backup codes, only regeneration
* refactor: Ensure secure fields are deleted from user data in getUserController
- Deleted setBalanceConfig middleware and its associated file.
- Introduced createSetBalanceConfig factory function to create middleware for synchronizing user balance settings.
- Updated auth and oauth routes to use the new balance configuration middleware.
- Added comprehensive tests for the new balance middleware functionality.
- Updated package versions and dependencies in package.json and package-lock.json.
- Added balance types and updated middleware index to export new balance middleware.
