v1.9.0

What's Changed

• fix(videos): allow up to 9 reference images on Seedance 2.0 by @steebchen in #3027
• chore(deps): bump better-auth from 1.6.18 to 1.6.23 in the better-auth group across 1 directory by @dependabot[bot] in #3014
• feat(compliance): provider headquarters filter by @steebchen in #3018
• feat: gate provider listing behind verified email by @steebchen in #3029
• feat: per-mapping reasoning efforts, forward as-is by @steebchen in #3028
• fix: reset stale reasoning effort, clarify docs by @steebchen in #3030
• feat(ui): expandable details on model mapping card by @steebchen in #3031
• fix: prevent duplicate Stripe customers in top-ups by @steebchen in #3033
• feat(devpass): add raw revenue and cumulative view by @steebchen in #3035
• fix: idempotency keys for Stripe customer creation by @steebchen in #3036
• feat: add rename action to API key options menu by @steebchen in #3038
• fix(ui): OG image logo fixes, spacexai redirects by @smakosh in #3040
• ci(e2e): add shard job failure detection by @steebchen in #3032
• fix(devpass): self-heal leaked tier-change claims by @steebchen in #3039
• feat: self-refund for credit and plan purchases by @steebchen in #2984
• fix(api): validate chat support body with zod by @steebchen in #3000
• docs: document the verbosity parameter by @steebchen in #3043
• fix(api): don't error-log bogus Stripe webhook signatures by @steebchen in #3001
• fix(gateway): flatten xiaomi tool result to string by @steebchen in #3045
• fix(gateway): 400 on malformed tool_call arguments by @steebchen in #2997
• feat: UI/UX polish for landing and DevPass by @smakosh in #3052
• fix: treat minimax abort finish reason as upstream error by @steebchen in #3053
• feat(models): declare reasoning efforts for xai by @AmineAce in #3046
• fix(api): route cached mutations through cdb by @serhiizghama in #2898
• feat: raise provider listing fee to $1000 by @steebchen in #3056
• feat(models): declare reasoning effort for gemini-2.5-flash by @AmineAce in #3048
• feat(models): declare reasoning effort for glm-5.2 by @AmineAce in #3049
• feat(routing): add filtered providers and stripped params to routing metadata by @steebchen in #1657
• perf(ui): improve cwv, a11y, seo on top pages by @smakosh in #3054
• feat: handle together-ai minimax-m2.7 deprecation by @steebchen in #3057
• feat(models): declare reasoning efforts for google-ai-studio by @AmineAce in #3047
• fix: skip drizzle update on empty PATCH body by @steebchen in #3058
• feat: gross revenue splits on admin dashboard by @steebchen in #3059
• fix(gateway): block provider routing on dev plans always by @steebchen in #2622
• feat(devpass): premium cap as % of credits by @steebchen in #2835
• fix(worker): count premium usage for prefixed usedModel by @steebchen in #3044
• fix: e2e failures — dead models, caps, vertex keys by @steebchen in #3012
• feat: add legal links to provider request form by @steebchen in #3064
• feat(models): add glm-5.2 via alibaba by @steebchen in #3067
• fix(gateway): map reasoning_effort to Moonshot thinking by @steebchen in #3065
• test: deflake cancellation spec with mock sync by @steebchen in #3068
• fix(admin): drop cents in dashboard metric cards by @steebchen in #3069
• feat(error-handler): handle upstream gateway errors by @steebchen in #3062
• fix: handle reasoning summary done streaming events by @steebchen in #3063
• fix(code): align premium fair-use copy by @steebchen in #3070
• feat(models): declare reasoning effort for deepseek by @AmineAce in #3072
• feat: allow hiding DevPass profile picture by @steebchen in #3073
• feat(devpass): default service tier setting by @steebchen in #3071
• docs: fix premium weekly cap window wording by @smakosh in #3074
• fix(api): validate vertex service-account keys by @steebchen in #3066
• feat(gateway): exact reset time in premium cap error by @smakosh in #3076
• docs: reflect full-charge upgrade model in copy by @steebchen in #3077
• feat: premium tier visibility + shared models dir by @smakosh in #3078
• feat(models): declare reasoning efforts for meta by @AmineAce in #3079
• feat(docs): expand Features section by default by @smakosh in #3080
• feat(admin): redesign dashboard overview by @smakosh in #3081
• feat(models): declare reasoning efforts for sakana fugu-ultra by @AmineAce in #3083
• feat: show disabled actions for playground api key by @steebchen in #3086
• feat(devpass): remove allow-all-models option by @steebchen in #3087
• fix(code): align DevPass invoice table columns by @steebchen in #3089
• fix: reasoning_effort for alibaba/minimax/xiaomi by @steebchen in #3075
• fix: novita deepseek-v3.2 reasoning suppression by @steebchen in #3091
• feat(models): add kimi-k3 by @smakosh in #3092
• feat(anthropic): add extended thinking support by @steebchen in #3088
• feat: add devpass toggle to revenue chart by @steebchen in #3095
• feat(models): declare reasoning efforts for bytedance seed models by @AmineAce in #3096
• fix(devpass): skip card dedupe in local dev by @steebchen in #3094
• fix: forward gateway 4xx errors in embedTexts by @steebchen in #3098
• feat(stealth-provider): redact error details in logs by @steebchen in #3090
• feat(devpass): add weekly premium Reset Passes by @smakosh in #3093
• feat(code): add 3D passport to DevPass profiles by @smakosh in #3099
• feat(devpass): notify Discord on Reset Pass buys by @smakosh in #3106
• fix(models): declare reasoning effort tiers for deepseek v4 by @steebchen in #3105
• fix(gateway): stop false-negative on large structured SSE JSON by @steebchen in #3107
• fix(gateway): route Bedrock models without a global profile by @steebchen in #3104
• fix(gateway): hash prompt_cache_key before forwarding by @steebchen in #3110
• fix(gateway): rewrite developer role for opted-out mappings by @steebchen in #3103
• fix(gateway): keep session-sticky retries on pinned provider by @steebchen in #3102
• fix(gateway): handle deprecated top_p/temp on Claude by @steebchen in #3108
• feat(code): redesign profile sharing, add LinkedIn by @smakosh in #3111
• feat: support LLM_OPENAI_BASE_URL env override by @steebchen in #3112
• feat: add reset pass sales KPI to admin devpass by @smakosh in #3115
• feat(admin): split reset pass revenue, 2-row KPIs by @smakosh in #3116
• feat(models): add service tiers for gpt-5.1/5.2 by @steebchen in #3117
• fix(docs): canonicalize shared guides to main site by @smakosh in #3118
• feat(code): gate reset passes on usage, add refunds by @smakosh in #3120
• fix(alibaba): declare reasoningMaxTokens for kimi-k2.5 by @AmineAce in #3113
• feat(gateway): add /v1/key plan status endpoint by @smakosh in #3123
• fix(reset-pass): prevent double-refunding by @steebchen in #3124
• feat(models): declare reasoning efforts for xai grok-4.3 on aws-bedrock by @AmineAce in #3125
• feat(novita): add kimi-k3 mapping by @steebchen in #3122
• feat(admin): show error classification in mappings by @steebchen in #3126
• test: harden cancellation spec log waits by @steebchen in #3128
• feat: ignore expected errors in unstable mappings by @steebchen in #3129
• feat(devpass): revamp coding models surfaces by @smakosh in #3132
• feat(models): declare reasoning efforts for deepseek-v4-pro by @AmineAce in #3131
• chore(deps-dev): bump commitlint in / by @dependabot[bot] in #3135
• chore(deps-dev): bump eslint-import-resolver-typescript in / by @dependabot[bot] in #3137
• fix(bedrock): drop empty messages in Converse builder by @steebchen in #3127
• fix(models): correct gpt-5.1 reasoning effort tiers by @steebchen in #3101
• feat(models): declare reasoning efforts for glm-5.2 on bytedance by @AmineAce in #3138
• fix(moderations): retry with next env key on auth error by @steebchen in #3130
• test: cover each declared reasoning effort in e2e by @steebchen in #3139
• fix(models): disable vision on tundra kimi-k2.6 by @steebchen in #3140
• fix: retry 405 responses with another key by @steebchen in #3141
• feat(models): add Gemini TTS on Vertex+AI Studio by @steebchen in #3143
• chore(deps-dev): bump @types/node in / by @dependabot[bot] in #3134
• feat: group unstable mapping errors by streaming by @steebchen in #3144
• chore: gitignore .empryo local tooling directory by @AmineAce in #3146
• feat: rollover credits and timing choice on upgrade by @steebchen in #3147
• docs(blog): add kimi k3 SEO cluster (5 posts) by @smakosh in #3150

Full Changelog: v1.8.0...v1.9.0