ptr.pet / cliproxyapi
sloproxy
Star
Fork 2
Watch

Commits on main

fix(claude): stop cloaking instructions as conversation turns main
nekomimi.pet committed
fix(claude): keep rebuilt system instructions top-level
nekomimi.pet committed
merge upstream main with local compatibility fixes
nekomimi.pet committed
chore(codex): update codex user-agent to 0.153.3
Luis Pater committed
feat(usage): propagate session and parent session hierarchy to usage reporting queue
sususu committed
fix(codex): scope usage limit errors to credentials and parse flexible quota resets
Luis Pater committed
fix(claude): avoid mid-conversation system splicing for advisor calls or results
Luis Pater committed
fix(watcher): preserve concurrent file updates during auth snapshot rescans
Luis Pater committed
fix(aistudio): normalize thinking level to uppercase
Luis Pater committed
chore(models): remove gpt-5.4 and gpt-5.4-mini models
Luis Pater committed
fix(codex): preserve empty supported_reasoning_levels array
Luis Pater committed
feat(models): add gpt-6-astra model and update codex client configurations
Luis Pater committed
perf(antigravity): batch reasoning replay mutations
rome-xi committed
fix(gemini): ensure functionResponse normalizes to user role in Gemini request normalizer
sususu committed
fix(antigravity): use ContentHasGeminiFunctionResponse instead of gjson projection
sususu committed
fix(translator): flush pending developer notice before intervening user turn
sususu committed
fix(translator): preserve Gemini prompt cache by demoting mid-session developer messages (#5490)
sususu committed
feat(claude): harden probe and helper request classification, diagnostics isolation, and late cloaking
sususu committed
feat(claude): add Fable 5.1 reporting outcomes block and post-payload reconciliation
sususu committed
feat(claude): add 2.1.258 dynamic beta headers, model fallbacks, and paired cache TTL
sususu committed
feat(claude): implement 2.1.258 billing header fingerprint chain and upstream request continuity
sususu committed
fix(antigravity): tighten replacement offset guard in tool provenance degradation
sususu committed
fix(auth): preserve concurrent modifications during auth refresh and preparation
Luis Pater committed
fix(translator/claude): downgrade strict mode when schema misses required properties
Luis Pater committed
fix(translator/interactions): avoid tool name collisions with Antigravity intrinsic tools
Luis Pater committed
Merge pull request #5435 from huangruiteng/codex/openai-compat-bounded-rate-limit-waits
Luis Pater committed
Merge pull request #5452 from deathemperor/docs/add-infinitus
Luis Pater committed
feat(plugin): omit stream chunk history on payload chunks for schema v5
Luis Pater committed
fix(translator/claude): handle incomplete status and terminal state on max_tokens
Luis Pater committed
perf(antigravity): batch replay degradation rewrites (#5461)
rome-xi committed
fix(translator/claude): defer message_delta and cache streaming usage
Luis Pater committed
fix(translator/gemini): nest image parts inside functionResponse
Luis Pater committed
feat(openai): send periodic ping control frames during responses websocket streaming
Luis Pater committed
feat(codex): support orphan delegation compatibility via orphan-delegation-compatibility
Luis Pater committed
feat: add Swiftproxy Sponser
Supra4E8C committed
feat: 添加 AxisNow 赞助信息及相关图像到 README 文件
Supra4E8C committed
chore: remove sponsorship images for Claude API and Code0 from README files
Supra4E8C committed
feat(session): derive distinct branch session ID, parent lineage on Merkle LCP forks, and enhance Codex fork/subagent affinity (#5418) (#5454)
sususu98 committed
fix(auth): prevent individual model quota cooldowns from blocking credential
Luis Pater committed
docs: add Infinitus to the projects based on CLIProxyAPI
deathemperor committed
feat(routing): make subagent session affinity configurable via session-affinity-subagents
Luis Pater committed
Merge pull request #5436 from router-for-me/models
Luis Pater committed
fix(codex): clear unsupported reasoning levels
hkfires committed
refactor(kimi): use request thinking helper
hkfires committed
feat(auth): use home model capabilities for thinking
hkfires committed
fix(models): preserve home model capability metadata
hkfires committed
fix(openai-compat): honor bounded rate-limit waits
huangruiteng committed
fix(claude): upgrade default Claude Code baseline and fingerprint to 2.1.258
sususu committed
fix(auth): enforce minimum cooldown floor and track attempted credentials on 429
Luis Pater committed
fix(home): map user_credits_insufficient to 402 and user_period_limit_exceeded to 429
Luis Pater committed
fix(auth): validate access token expiration and retain valid credentials on refresh failure
Luis Pater committed
chore(registry): remove "minimal" level from dynamic_allowed definitions in models
Luis Pater committed
feat(registry): add claude fable 5.1 and gemini 3.8 flash models
Luis Pater committed
fix(antigravity): bypass quota cooldowns and credit hints when cooling is disabled
Luis Pater committed
fix(antigravity): strip tool config, labels, and session id in token counting
Luis Pater committed
fix: harden concurrent session handling, listener lifecycle, and token accounting
Luis Pater committed
fix(antigravity): preserve unsigned gemini thinking blocks with trailing carriers
Luis Pater committed
fix(antigravity): retry model fetching per endpoint and prioritize daily base url
Luis Pater committed
fix(antigravity): support native thinking signatures without prefixes in claude translator
Luis Pater committed
feat(translator): support cache write tokens in claude responses
Luis Pater committed