Release v0.4.9

This release introduces critical updates to proxy configuration and reliability, with structural adjustments.

New Features

• Added configuration files for enhanced settings.
• Introduced new handler implementations.

Bug Fixes

• Fixed availability gate handling issues.
• Simplified retry logic in availability tests.

Improvements

• Optimized performance metrics.
• Enhanced documentation clarity.

Documentation

• Updated README and internal guides.

Chores

• Updated CI/CD pipelines.
• Refactored dependency management.

Release v0.4.8

Summary
This release introduces a robust Windows autostart fix and a built‑in self‑update mechanism, enabling users to keep routatic‑proxy current directly from the command line. Enhancements to autostart handling, command‑line formatting, and background process logging improve reliability on Windows, while the new updater package adds version comparison, checksum verification, and GitHub API integration.

---

New Features

• Add Windows autostart fix – handle WindowsApps aliases and improve command‑line argument formatting.
• Introduce self‑update command – check for and apply updates from GitHub.
• Create updater package – download, verify, and install updates with checksum validation.
• Enhance User‑Agent handling – set a custom User‑Agent for GitHub API requests.

Bug Fixes

• Update .gitignore – include .kimchi and remove obsolete documentation files.
• Replace ping with timeout – improve backup deletion reliability on Windows.

Improvements

• Autostart status command – now checks if the server process is running and logs startup errors.
• Background process startup – add error handling and logging improvements.
• Windows command‑line building – unit tests added for path extraction and argument formatting.
• Updater functionality – tests added for version comparison and checksum verification.

Documentation

• Add Windows autostart specification – new documentation file detailing the fix.
• Update INSTALLATION.md & README.md – reflect new features and usage.

Chores

• Add cmd/routatic-proxy/update.go – command implementation for updates.
• Update internal daemon files – autostart_windows.go, background.go, and their tests.
• Add updater source and tests – internal/updater package and corresponding test files.
• Update .gitignore – maintain clean repository state.

Release v0.4.7

User Safety: safe

Release v0.4.6

This release focuses on expanding the project's accessibility through comprehensive localization. Significant efforts were made to provide full Chinese documentation for core setup and operational processes.

Documentation

• Add Chinese translation for the installation guide
• Add Chinese translation for the configuration guide
• Add Chinese translation for the models documentation
• Add a new Chinese troubleshooting guide
• Update the main README with Chinese language support
• Improve existing README content and structure

Release v0.4.5

This release ensures seamless macOS compatibility and performance enhancements.

New Features

• macOS system tray and GUI webview integration
• Configuration refinements for custom proxy settings

Bug Fixes

• Proxy restart synchronization issues
• GUI input handling improvements

Improvements

• Performance optimizations for large-scale deployments
• Enhanced security protocols for sensitive data

Documentation

• Updated README with bilingual guidance
• Configuration documentation additions

Chores

• Automated build pipeline updates
• Dependency version management

Breaking Changes
⚠️ Breaking changes include GUI layout adjustments requiring UI adjustments.

Release v0.4.4

Breaking Changes 🚨

• Update configuration to use provider‑specific API keys; existing configs must be migrated.

New Features

• Add support for provider‑specific API keys with comma‑separated values.
• Implement fallback strategies when usage limits are exceeded.
• Enhance environment variable handling for API key configuration.

Bug Fixes

• Fix handling of usage limit errors to prevent unintended fallback.
• Add validation for provider‑specific API keys in the configuration.
• Adjust formatting of API key configuration tests.
• Remove obsolete routatic‑hosted binary and update .gitignore.

Improvements

• Strengthen provider‑specific API key handling and increase test coverage.
• Refine error handling and fallback logic for better reliability.

Documentation

• Update CLAUDE.md with new configuration details and usage examples.

Chores

• Add routatic-hosted to .gitignore.
• Clean up test artifacts and streamline CI setup.

Release v0.4.3

⚠️ Breaking Changes

• Vision routing now only triggers when the latest user message contains a new (unseen) image; historic images no longer cause automatic switches to vision models.
• Some models previously marked as vision‑capable have been reverted; they will no longer be used for image input unless explicitly re‑verified.

New Features

• Add Images field and NormalizedImage struct to preserve media type and base64 data across normalization.
• Extend model registry with three 1M‑context vision‑capable models: qwen3.7-max, qwen3.7-plus, mimo-v2-omni.
• Flip Vision: true for mimo-v2.5-pro, mimo-v2.5, and minimax-m3 based on empirical verification.
• Update recommended vision scenarios: vision → kimi-k2.7-code, vision_complex → mimo-v2-omni, vision_long_context → qwen3.7-plus.

Bug Fixes

• Preserve image data in the normalize/denormalize bridge by checking for non‑empty source data.
• Fix request routing to trigger vision only on new images, eliminating false positives from stale history keywords.
• Correct over‑eager vision flag for unsupported models (minimax-m3, qwen3.7-max).
• Streamline request facts by removing unused fields from scenario logic.

Improvements

• Enhance end‑to‑end reliability of image handling for 32×32 PNG test cases.
• Refine model metadata updates to avoid speculative vision flag changes.

Documentation

• Update internal/config/model_registry.go to reflect new vision‑capable entries.
• Adjust scenario test expectations to match the revised routing behavior.

Chores

• Update internal imports and struct definitions to accommodate new Images field.
• Clean up unused imports and dead code in normalization and scenario packages.

Release v0.4.2

Breaking Changes

• ⚠️ No breaking changes in this release.

New Features

• Add regression test for small‑max_tokens scenario.

Bug Fixes

• Fix capacity filter to honor small max_tokens and skip models only when their context window is exhausted below the output floor.

Improvements

• Enhance capacity filter logic to correctly evaluate output token thresholds.

Documentation

• No documentation updates.

Chores

• Update .gitignore to include brag-output and .DS_Store.

Release v0.4.1

Release Notes

This release focuses on expanding OpenCode model support and improving the codebase architecture. New models including GLM-5.2, Kimi K2.7 Code, and Qwen3.7 variants are now available, while deprecated models have been documented. The model classification logic has been refactored into a dedicated package for better maintainability.

⚠️ Breaking Changes

• Deprecated several older OpenCode models - update your configurations to use supported alternatives

New Features

• Add support for new OpenCode Go models: GLM-5.2, Kimi K2.7 Code, Qwen3.7 Plus, and Qwen3.7 Max
• Refactor model classification logic into a new internal/models package for improved code organization

Bug Fixes

• Update model ID references to mimo-v2.5-pro across configurations and tests for consistency

Improvements

• Update model metadata with new context windows and output token limits
• Enhance routing and classification logic in opencode.go to properly handle new models
• Update default model ID to glm-5.1 in configuration files

Documentation

• Document new models and deprecated models in README.md and MODELS.md
• Update endpoint classification documentation for model integration
• Refresh architecture.md and howto-add-model.md with current implementation details

Chores

• Add unit tests for new models and update existing tests to ensure proper routing and classification
• Update example configuration in config.example.json to reflect new model defaults

Release v0.4.0

User Safety: safe