v0.1.3

Changelog

• 9acb709 Address latest review findings
• b8641fc Address review feedback on perf coverage
• b50dbc9 Ensure codex completed events always publish usage stats
• 5197cf0 Ensure codex streaming requests force stream mode
• cf47550 Fix codex completed-event reader and stress test gating
• 6487db8 Fix codex request cache and SSE drain regressions
• c4ff5e8 Fixed: router-for-me#2274
• db66efc Force stream mode in codex streaming fallback path
• 31026cd Optimize auth hot paths and add TTFT stress coverage
• cfc0e3f Optimize codex request preparation under extreme load
• bb88165 build: add freebsd support to releases
• d90bded docs(readme): update QQ group information in Chinese README
• 37c8bab feat(api): support batch auth file upload and delete
• 180a7ed feat(codex): pass through codex client identity headers
• 3a6ddfd feat(openai-compat): add per-model thinking support
• 398b0f5 fix(auth): restore filestore project id hydration
• 839c863 fix(openai-compat): preserve legacy thinking passthrough
• 6781f8e fix(security): harden management panel asset updater
• 2d897a7 fix(usage): persist stats across config changes
• 6fa038a fix: finalize upstream v6.9.3 backport integration
• 869530b fix: keep a fallback turn for system-only Claude inputs
• 60a7fdb fix: map OpenAI system messages to Claude top-level system
• b86b0c7 fix: reject oversized downloads instead of truncating; warn on unverified fallback
• e7e1ca5 refactor(config): replace auto-update-panel with disable-auto-update-panel for clarity
• ddf062c 持久化使用统计并支持自动恢复

v0.1.2

Changelog

• 61ab025 Backport safe upstream fixes for PR 8
• abaecb2 Expand safe-subset coverage and benchmarks
• f438e59 Handle refresh 401 in auth maintenance
• 1162926 Handle websocket model rewrite errors
• 39632d7 Normalize compact Codex instructions field
• ee4a40a Optimize codex builtin tool normalization fast path
• 6aa81ac feat(claude): add switch for device profile stabilization
• 4e35f96 fix(auth): avoid blocking oauth callback prompt
• 2bd7c74 fix(auth): avoid redundant scheduler index rebuild on lazy shards
• fa33d8e fix(auth): carry safe scheduler and proxy follow-ups
• 8ebb93d fix(auth): stop cross-credential retries on invalid requests
• 8ca4c20 fix(claude): avoid racy fingerprint downgrades
• ae1d7e9 fix(claude): keep configured baseline above older fingerprints
• e8aaca7 fix(claude): learn official fingerprints after custom baselines
• 74730aa fix(claude): pin stabilized OS arch to baseline
• c7b5907 fix(claude): preserve legacy user agent overrides
• ed0e297 fix(claude): refresh cached fingerprint after baseline upgrades
• f36a579 fix(claude): restore legacy runtime OS arch fallback
• f925950 fix(codex): strip websocket model prefixes upstream
• ea217b0 fix(management): clamp invalid retry config
• c248504 fix: backport translator and auth fallback fixes
• 6f02529 fix: stabilize claude device fingerprint
• 03f6281 perf(auth): harden scheduler fast paths under load

v0.1.1

Changelog

• 92ba296 ci(workflow): publish beta from main and stable from tags
• d34cdd0 perf(codex): fast-path responses request translation

v0.1.0

Changelog

• 6f9c23a router-for-me#167
• 7063a17 router-for-me#293
• 34855bc Fix model switch logic when quota is exceeded
• 56c8297 Handle data: without trailing space in streaming responses
• cbf9a57 build(goreleaser): set CGO_ENABLED=0 for cli-proxy-api binaries
• 257621c chore(executor): update default agent version and simplify const formatting
• 8b13c91 docs(internal): add Codex instruction guides for GPT-5 CLI
• 93fa1d1 docs: add Amp CLI integration guide to Chinese documentation
• 6a2906e feat(antigravity): add support for Claude-Opus-4-5-Thinking model
• 76af454 feat(antigravity): enhance handling of "thinking" content and refine Claude model response processing
• 361443d feat(api): add GetLatestVersion endpoint to fetch latest release version from GitHub
• cc3cf09 feat(auth): add AuthIndex for diagnostics and ensure usage recording
• 7757210 feat(auth): implement Antigravity OAuth authentication flow
• 2d84d2f feat(auth, executor, cmd): add Antigravity provider integration
• d114d8d feat(config): add TLS support for HTTPS server configuration
• 8d23ffc feat(executor): add model alias mapping and improve Antigravity payload handling
• ac06438 feat(executor, translator): enhance token handling and payload processing
• aa57f32 feat(instructions): add detailed agent behavior guidelines for Codex CLI
• e308288 feat(logging, middleware): add error-based logging support and error log management
• bf116b6 feat(registry): add GPT-5.1 Codex Max model definitions and support
• ee6429c feat(registry): add Gemini 3 Pro Image Preview model and remove Claude Sonnet 4.5 Thinking
• 17016ae feat(registry): add Gemini 3 Pro Preview model definition
• 01b7b60 feat(registry): add Gemini 3 Pro Preview model definition
• d2e4639 feat(registry): add context length and update max tokens for Claude model configurations
• 371324c feat(registry): expand Gemini model definitions and support Vertex AI
• fcd98f4 feat(runtime): add payload configuration support for executors
• cf9b9be feat(runtime): extend executor support for GPT-5.1 Codex and variants
• c1031e2 feat(translator): add Antigravity translation logic
• 8977121 feat(translator): add ThoughtSignature handling in Gemini request transformations
• 4eab141 feat(translator): add support for reasoning/thinking content blocks in OpenAI-Claude and Gemini responses
• 9d50a68 feat(translator): improve content processing and Antigravity request conversion
• 32d3809 feat(util): add -reasoning suffix support for Gemini models
• d6352dd feat(util): add DeleteKey function and update antigravity executor for Claude model compatibility
• 4307f08 feat(watcher): optimize auth file handling with hash-based change detection
• f0711be fix(auth): prevent access to removed credentials lingering in memory
• bb9955e fix(auth): resolve index reassignment issue during auth management
• 9875565 fix(claude translator): ensure default token counts when usage data is missing
• 52c17f0 fix(executor): comment out redundant code for setting Claude system instructions
• 113db3c fix(executor): update antigravity executor to enhance model metadata handling
• a748e93 fix(executor, auth): ensure index assignment consistency for auth objects
• 506f111 fix(handlers): refactor API response capture to append data safely
• 23a7633 fix(registry): update Thinking parameters and replace Gemini-3 Preview with Gemini-2.5 Flash Lite
• db2d22c fix(runtime): simplify scanner buffer allocation in executor implementations
• 0a47b45 fix(translator): add conditional check for key renaming in Gemini tools
• 261f08a fix(translator): adjust key renaming logic in Gemini request processing
• 9f95b31 fix(translator): enhance handling of mixed output content in Claude requests
• c8cee54 fix(translator): ensure partial content is retained while skipping encrypted thoughtSignature
• 7a8e00f fix(translator): handle missing parameters in Gemini tool schema gracefully
• 346b663 fix(translator): handle non-JSON output gracefully in function call outputs
• 72c7ef7 fix(translator): handle non-JSON output parsing for OpenAI function responses
• 39621a0 fix(translator): normalize function calls and outputs for consistent input processing
• 0bcae68 fix(translator): preserve raw JSON encoding in function call outputs
• ddb0c0e fix(translator): reintroduce thoughtSignature bypass logic for model parts
• 41ee444 fix(translator): rename responseSchema key for generationConfig
• 1c815c5 fix(translator): simplify string handling in Gemini responses
• d4d5298 refactor(antigravity): handle anyOf property, remove exclusiveMinimum, and comment unused prod URL
• 1d0f030 refactor(api/config): centralize legacy OpenAI compatibility key migration
• 327cc70 refactor(auth): use customizable HTTP client for Antigravity requests
• 0fd2abb refactor(cliproxy, config): remove vertex-compat flow, streamline Vertex API key handling
• d50b0f7 refactor(executor): simplify Gemini CLI execution and remove internal retry logic
• db81331 refactor(middleware): extract request logging logic and optimize condition checks
• 0586da9 refactor(registry): move Gemini 3 Pro Preview model definition to base set
• 1434bc3 refactor(registry): remove Qwen3-Coder from model definitions
• 1ccb016 refactor(runtime): centralize reasoning effort logic for GPT models
• 98596c0 refactor(translator): remove service_tier from Codex OpenAI request payload
• 09970dc Accept Geminis Review Suggestion
• 8193392 Add AMP fallback proxy and shared Gemini normalization
• 4b2d40b Add CLIProxyAPI Dashboard to 'Who is with us?' section
• 13bb7cf Add CLIProxyAPI Tray information to README
• 8bce696 Add CLIProxyAPI Tray section to README_CN.md
• 889b2ab Add CPA auth maintenance queue and concurrency hardening
• c86545d Add Chinese README and update project files
• 963a095 Add Claude Proxy VSCode extension to README
• 3b484ae Add Claude Proxy VSCode to README_CN.md
• 3c0c61a Add Claude compatibility and enhance API handling
• dca8d5d Add CodMate app information to README
• 2a7fd1e Add CodMate description to README_CN.md
• c62e45e Add Codex API key support and Gemini 2.5 Flash-Lite model documentation updates
• 67e0b71 Add Codex load balancing documentation and refine JSON handling logic
• 70988d3 Add Codex websocket header defaults
• 9b627a9 Add Docker Compose
• 3687963 Add Docker support with CI/CD workflow and usage instructions
• 2406cc7 Add GPT-5 Codex model support and configuration options in documentation
• 3e09bc9 Add Gemini 2.5 Flash-Lite Model
• b3607d3 Add Gemini-compatible API and improve error handling
• c0eaa41 Add Gemini-to-Gemini request normalization and passthrough support
• 6164f5c Add JSON annotations to configuration structs and new /config management endpoint
• 7e0fd1e Add Keep-Alive header
• f5f26f0 Add Kimi (Moonshot AI) provider support
• 545dff8...