@bubblebrain-ai/bubble 0.0.63 → 0.0.65

package/dist/model-catalog.js

@@ -0,0 +1,399 @@
+import { OPENROUTER_MODEL_ID } from "./provider-model-policy.js";
+export const BUILTIN_PROVIDERS = [
+    { id: "openrouter", name: "OpenRouter", baseURL: "https://openrouter.ai/api/v1" },
+    { id: "openai", name: "OpenAI", baseURL: "https://api.openai.com/v1", supportsOAuth: true },
+    { id: "openai-codex", name: "OpenAI Codex (ChatGPT)", baseURL: "https://chatgpt.com/backend-api" },
+    { id: "opencode-zen", name: "OpenCode Zen", baseURL: "https://opencode.ai/zen/v1", protocol: "openai-responses" },
+    // Grok subscription models via the CLI chat proxy. OAuth-only: the proxy
+    // accepts xAI session bearers (from /login grok), not api.x.ai API keys.
+    { id: "grok", name: "Grok Subscription", baseURL: "https://cli-chat-proxy.grok.com/v1", supportsOAuth: true },
+    { id: "anthropic", name: "Anthropic", baseURL: "https://api.anthropic.com", protocol: "anthropic-messages" },
+    { id: "deepseek", name: "DeepSeek", baseURL: "https://api.deepseek.com" },
+    // Native Gemini API via the AI SDK google provider. Users who configured the
+    // old OpenAI-compat endpoint can keep it by setting protocol "openai-chat"
+    // and the /openai baseURL explicitly in models.json.
+    { id: "google", name: "Google Gemini", baseURL: "https://generativelanguage.googleapis.com/v1beta", protocol: "ai-sdk" },
+    { id: "zhipuai", name: "Zhipu AI", baseURL: "https://open.bigmodel.cn/api/paas/v4" },
+    { id: "zhipuai-coding-plan", name: "Zhipu AI Coding Plan", baseURL: "https://open.bigmodel.cn/api/coding/paas/v4" },
+    { id: "zai", name: "Z.AI", baseURL: "https://api.z.ai/api/paas/v4" },
+    { id: "zai-coding-plan", name: "Z.AI Coding Plan", baseURL: "https://api.z.ai/api/coding/paas/v4" },
+    { id: "alibaba", name: "Alibaba DashScope", baseURL: "https://dashscope.aliyuncs.com/compatible-mode/v1" },
+    // Bailian token plan: its own endpoint and quota, separate from the
+    // pay-per-token DashScope provider above. Hosts several vendors' models.
+    { id: "bailian-token-plan", name: "阿里云百炼 Token Plan", baseURL: "https://token-plan.cn-beijing.maas.aliyuncs.com/compatible-mode/v1" },
+    { id: "doubao", name: "Doubao (Volcengine Ark)", baseURL: "https://ark.cn-beijing.volces.com/api/v3", protocol: "ark-responses" },
+    { id: "minimax", name: "MiniMax Token Plan", baseURL: "https://api.minimaxi.com/anthropic", protocol: "anthropic-messages" },
+    { id: "minimax-anthropic", name: "MiniMax API", baseURL: "https://api.minimaxi.com/anthropic", protocol: "anthropic-messages" },
+    { id: "mimo-token-plan", name: "Xiaomi MiMo Token Plan", baseURL: "https://token-plan-cn.xiaomimimo.com/anthropic", protocol: "anthropic-messages" },
+    { id: "stepfun", name: "StepFun Step Plan", baseURL: "https://api.stepfun.com/step_plan/v1" },
+    { id: "stepfun-api", name: "StepFun API", baseURL: "https://api.stepfun.com/v1", protocol: "openai-chat" },
+    { id: "moonshot-cn", name: "Moonshot (国内 platform.moonshot.cn)", baseURL: "https://api.moonshot.cn/v1" },
+    { id: "moonshot-intl", name: "Moonshot (海外 platform.moonshot.ai)", baseURL: "https://api.moonshot.ai/v1" },
+    { id: "kimi-for-coding", name: "Kimi for Coding", baseURL: "https://api.kimi.com/coding/v1" },
+    { id: "groq", name: "Groq", baseURL: "https://api.groq.com/openai/v1" },
+    { id: "together", name: "Together AI", baseURL: "https://api.together.xyz/v1" },
+    { id: "fireworks", name: "Fireworks", baseURL: "https://api.fireworks.ai/inference/v1" },
+    { id: "local", name: "Local (OpenAI-compatible)", baseURL: "http://localhost:11434/v1" },
+];
+const ALL_OPENAI_LEVELS = ["off", "minimal", "low", "medium", "high", "xhigh"];
+const GPT56_LEVELS = ["low", "medium", "high", "xhigh", "max", "ultra"];
+const GPT56_LUNA_LEVELS = ["low", "medium", "high", "xhigh", "max"];
+const GPT51_LEVELS = ["off", "low", "medium", "high"];
+const GPT51_CODEX_MAX_LEVELS = ["off", "low", "medium", "high", "xhigh"];
+const GPT51_CODEX_MINI_LEVELS = ["off", "medium", "high"];
+const OPENAI_CHAT_LEVELS = ["off"];
+const MUSE_SPARK_LEVELS = ["minimal", "low", "medium", "high", "xhigh"];
+// Ox Alpha declares mandatory reasoning with exactly low/high/max through
+// OpenRouter's live model metadata. The endpoint rejects disabled reasoning.
+const OPENROUTER_OX_REASONING_LEVELS = ["low", "high", "max"];
+// Internal representation for APIs that expose thinking as enabled/disabled.
+// UI code must render supported toggle models as on/off, not as "medium" effort.
+const TOGGLE_THINKING_LEVELS = ["off", "medium"];
+// GLM-5.2 accepts OpenAI-style `reasoning_effort`. We expose high and max plus
+// "off", which disables thinking outright via `thinking: {type: "disabled"}`.
+const GLM_5_2_LEVELS = ["high", "max", "off"];
+// The GLM-5.3 family normalizes efforts into exactly three real tiers. Its
+// default is max, and disabling thinking maps to low rather than turning it off.
+const GLM_5_3_LEVELS = ["low", "high", "max"];
+// Kimi K2.7 Code variants only support thinking mode (disabling it errors), so
+// "off" is not offered — the model is always in its thinking variant.
+const KIMI_THINKING_ONLY_LEVELS = ["medium"];
+// Kimi K3 takes OpenAI-style reasoning_effort (low/high/max — no "medium"),
+// verified against the live endpoint: low spent 92 reasoning tokens on a task
+// where high spent 1310. Thinking is disableable on k3 (thinking.type
+// "disabled"); the 256k variant exposes effort only, so it has no "off".
+// K3.1 Pro (and its highspeed slot) is effort-only too: GET /coding/v1/models
+// reports reasoning.type "only" and the endpoint 400s on thinking disabled.
+// The default of "high" is Bubble's choice and matches the coding endpoint's
+// reported default; Moonshot's open platform defaults to max server-side.
+const KIMI_K3_LEVELS = ["off", "low", "high", "max"];
+const KIMI_K3_EFFORT_ONLY_LEVELS = ["low", "high", "max"];
+// DashScope's Moonshot passthrough (kimi/ ids) documents K3 as thinking-only
+// with reasoning_effort "max" as its sole value. Not live-verified: the probing
+// account has the passthrough product disabled.
+const KIMI_PASSTHROUGH_K3_LEVELS = ["max"];
+// DeepSeek V4 defaults to thinking/high, but the OpenAI-compatible API also
+// supports disabling thinking and selecting low/high/max effort explicitly.
+const DEEPSEEK_V4_LEVELS = ["off", "low", "high", "max"];
+// Bailian token plan grades, read off the endpoint's own 400 responses
+// ("'reasoning_effort' must be one of: ..."), which differ per model. "none"
+// maps to our "off". qwen3.6-flash / qwen3.7-* stop at xhigh; deepseek-v4-pro
+// there has no off switch.
+const BAILIAN_FULL_LEVELS = ["off", "minimal", "low", "medium", "high", "xhigh", "max"];
+const BAILIAN_NO_MAX_LEVELS = ["off", "minimal", "low", "medium", "high", "xhigh"];
+const BAILIAN_DEEPSEEK_LEVELS = ["low", "medium", "high", "xhigh", "max"];
+const STEPFUN_REASONING_LEVELS = ["off", "low", "medium", "high"];
+const DOUBAO_SEED_REASONING_LEVELS = ["minimal", "low", "medium", "high"];
+const MINIMAX_M3_REASONING_LEVELS = ["off", "medium"];
+const MINIMAX_REASONING_LEVELS = ["medium"];
+// Anthropic exposes reasoning depth through output_config.effort (low | medium
+// | high | xhigh | max), not a token budget. Sonnet 5 also supports xhigh.
+// Fable 5+ and Opus 5.5 have thinking always on, so they have no "off".
+const ANTHROPIC_OPUS_EFFORT_LEVELS = ["off", "low", "medium", "high", "xhigh", "max"];
+const ANTHROPIC_SONNET_EFFORT_LEVELS = ["off", "low", "medium", "high", "max"];
+const ANTHROPIC_FABLE_EFFORT_LEVELS = ["low", "medium", "high", "xhigh", "max"];
+const ANTHROPIC_CHAT_LEVELS = ["off"];
+// Grok 4.5 exposes low/medium/high effort with no "off" (thinking always on);
+// the composer model declares no effort support at all.
+const GROK_45_LEVELS = ["low", "medium", "high"];
+// Gemini 4 documents only low/medium (default)/high. Verified live on
+// 2026-10-13: thinkingLevel=minimal returns 400, so there is no off/minimal.
+const GEMINI_4_LEVELS = ["low", "medium", "high"];
+// Gemini 4 Hydrogen adds xhigh above the Gemini 4 ladder and defaults to high.
+// Verified live on 2026-11-12: xhigh returns 200, minimal still returns 400
+// (Argon rejects xhigh).
+const GEMINI_4_HYDROGEN_LEVELS = ["low", "medium", "high", "xhigh"];
+const GEMINI_3_LEVELS = ["low", "medium", "high"];
+const GEMINI_3_FLASH_LEVELS = ["minimal", "low", "medium", "high"];
+// Verified against the native API on 2026-09-02: Gemini 3.7/3.8 Flash reject
+// thinkingLevel=minimal but accept thinkingBudget=0 plus low/medium/high.
+const GEMINI_37_38_FLASH_LEVELS = ["off", "low", "medium", "high"];
+const GEMINI_25_PRO_LEVELS = ["low", "medium", "high"];
+const GEMINI_25_FLASH_LEVELS = ["off", "low", "medium", "high"];
+export const BUILTIN_MODELS = [
+    // OpenRouter exposes a very large catalog, but Bubble intentionally offers
+    // only Ox Alpha. Discovery refreshes this entry's public metadata without
+    // expanding provider membership.
+    { id: OPENROUTER_MODEL_ID, name: "Ox Alpha", providerId: "openrouter", tier: "strong", reasoningLevels: OPENROUTER_OX_REASONING_LEVELS, defaultReasoningLevel: "max", contextWindow: 1048576 },
+    // Verified against /codex/models on 2026-09-17: same request shape as gpt-5.6-terra
+    // (responses-lite, six effort levels, 10k tool-output cap); minimal_client_version 0.153.0.
+    { id: "gpt-6-astra", name: "GPT-6-Astra", providerId: "openai-codex", tier: "strong", reasoningLevels: GPT56_LEVELS, defaultReasoningLevel: "medium", contextWindow: 372000, useResponsesLite: true, toolOutputTokenLimit: 10000 },
+    { id: "gpt-5.6-sol", name: "GPT-5.6-Sol", providerId: "openai-codex", tier: "balanced", reasoningLevels: GPT56_LEVELS, defaultReasoningLevel: "low", contextWindow: 372000, useResponsesLite: true, toolOutputTokenLimit: 10000 },
+    { id: "gpt-5.6-terra", name: "GPT-5.6-Terra", providerId: "openai-codex", tier: "strong", reasoningLevels: GPT56_LEVELS, defaultReasoningLevel: "medium", contextWindow: 372000, useResponsesLite: true, toolOutputTokenLimit: 10000 },
+    { id: "gpt-5.6-luna", name: "GPT-5.6-Luna", providerId: "openai-codex", tier: "strong", reasoningLevels: GPT56_LUNA_LEVELS, defaultReasoningLevel: "medium", contextWindow: 372000, useResponsesLite: true, toolOutputTokenLimit: 10000 },
+    { id: "gpt-5.5", name: "gpt-5.5", providerId: "openai-codex", tier: "strong", reasoningLevels: ALL_OPENAI_LEVELS, contextWindow: 272000, toolOutputTokenLimit: 10000 },
+    { id: "gpt-5.4", name: "gpt-5.4", providerId: "openai-codex", tier: "balanced", reasoningLevels: ALL_OPENAI_LEVELS, contextWindow: 272000 },
+    { id: "gpt-5.4-mini", name: "gpt-5.4-mini", providerId: "openai-codex", tier: "fast", reasoningLevels: ALL_OPENAI_LEVELS, contextWindow: 272000 },
+    { id: "gpt-5.3-codex", name: "gpt-5.3-codex", providerId: "openai-codex", reasoningLevels: ALL_OPENAI_LEVELS, contextWindow: 272000 },
+    { id: "gpt-5.3-codex-spark", name: "gpt-5.3-codex-spark", providerId: "openai-codex", reasoningLevels: ALL_OPENAI_LEVELS, contextWindow: 272000 },
+    { id: "gpt-5.2-codex", name: "gpt-5.2-codex", providerId: "openai-codex", reasoningLevels: ALL_OPENAI_LEVELS, contextWindow: 272000 },
+    { id: "gpt-5.2", name: "gpt-5.2", providerId: "openai-codex", reasoningLevels: ALL_OPENAI_LEVELS, contextWindow: 272000 },
+    { id: "gpt-5.1-codex-max", name: "gpt-5.1-codex-max", providerId: "openai-codex", reasoningLevels: GPT51_CODEX_MAX_LEVELS, contextWindow: 272000 },
+    { id: "gpt-5.1-codex-mini", name: "gpt-5.1-codex-mini", providerId: "openai-codex", tier: "fast", reasoningLevels: GPT51_CODEX_MINI_LEVELS, contextWindow: 272000 },
+    { id: "gpt-5.1", name: "gpt-5.1", providerId: "openai-codex", reasoningLevels: GPT51_LEVELS, contextWindow: 272000 },
+    { id: "grok-4.5", name: "Grok 4.5", providerId: "grok", tier: "strong", reasoningLevels: GROK_45_LEVELS, defaultReasoningLevel: "high", contextWindow: 500000 },
+    { id: "grok-composer-2.5-fast", name: "Grok Composer 2.5 Fast", providerId: "grok", tier: "fast", reasoningLevels: OPENAI_CHAT_LEVELS, contextWindow: 200000 },
+    { id: "gpt-4o", name: "gpt-4o", providerId: "openai", tier: "balanced", reasoningLevels: OPENAI_CHAT_LEVELS, contextWindow: 128000 },
+    { id: "gpt-4o-mini", name: "gpt-4o-mini", providerId: "openai", tier: "fast", reasoningLevels: OPENAI_CHAT_LEVELS, contextWindow: 128000 },
+    { id: "o1-preview", name: "o1-preview", providerId: "openai", tier: "strong", reasoningLevels: ["off", "low", "medium", "high"], contextWindow: 128000 },
+    { id: "o1-mini", name: "o1-mini", providerId: "openai", tier: "fast", reasoningLevels: ["off", "low", "medium", "high"], contextWindow: 128000 },
+    { id: "gpt-4-turbo", name: "gpt-4-turbo", providerId: "openai", reasoningLevels: OPENAI_CHAT_LEVELS, contextWindow: 128000 },
+    { id: "muse-spark-1.3-contributor-free", name: "Muse Spark 1.3 Free", providerId: "opencode-zen", reasoningLevels: MUSE_SPARK_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "muse-spark-1.2", name: "Muse Spark 1.2", providerId: "opencode-zen", reasoningLevels: MUSE_SPARK_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "muse-spark-1.2-contributor-free", name: "Muse Spark 1.2 Free", providerId: "opencode-zen", reasoningLevels: MUSE_SPARK_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    // Zen /chat/completions; models.dev/opencode metadata checked 2026-09-26.
+    { id: "space-bunny-free", name: "Space Bunny Free", providerId: "opencode-zen", reasoningLevels: ["low", "medium", "high", "xhigh", "max"], defaultReasoningLevel: "high", contextWindow: 1048576 },
+    // Current public lineup, verified against platform.claude.com on 2026-09-23.
+    { id: "claude-fable-5-1", name: "Claude Fable 5.1", providerId: "anthropic", tier: "strong", reasoningLevels: ANTHROPIC_FABLE_EFFORT_LEVELS, defaultReasoningLevel: "high", contextWindow: 1000000 },
+    { id: "claude-opus-5-5", name: "Claude Opus 5.5", providerId: "anthropic", tier: "strong", reasoningLevels: ANTHROPIC_FABLE_EFFORT_LEVELS, defaultReasoningLevel: "medium", contextWindow: 1000000 },
+    { id: "claude-sonnet-5", name: "Claude Sonnet 5", providerId: "anthropic", tier: "balanced", reasoningLevels: ANTHROPIC_OPUS_EFFORT_LEVELS, defaultReasoningLevel: "high", contextWindow: 1000000 },
+    { id: "claude-opus-5", name: "Claude Opus 5", providerId: "anthropic", tier: "strong", reasoningLevels: ANTHROPIC_OPUS_EFFORT_LEVELS, defaultReasoningLevel: "high", contextWindow: 1000000 },
+    { id: "claude-fable-5", name: "Claude Fable 5", providerId: "anthropic", tier: "strong", reasoningLevels: ANTHROPIC_FABLE_EFFORT_LEVELS, defaultReasoningLevel: "high", contextWindow: 1000000 },
+    { id: "claude-opus-4-8", name: "Claude Opus 4.8", providerId: "anthropic", tier: "strong", reasoningLevels: ANTHROPIC_OPUS_EFFORT_LEVELS, defaultReasoningLevel: "high", contextWindow: 1000000 },
+    { id: "claude-sonnet-4-6", name: "Claude Sonnet 4.6", providerId: "anthropic", tier: "balanced", reasoningLevels: ANTHROPIC_SONNET_EFFORT_LEVELS, defaultReasoningLevel: "high", contextWindow: 1000000 },
+    { id: "claude-haiku-4-5-20251001", name: "Claude Haiku 4.5", providerId: "anthropic", tier: "fast", reasoningLevels: ANTHROPIC_CHAT_LEVELS, contextWindow: 200000 },
+    { id: "deepseek-v4-flash", name: "deepseek-v4-flash", providerId: "deepseek", tier: "fast", reasoningLevels: DEEPSEEK_V4_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    // V4.1 Flash (2026-09-10): high is the default, not the only effort.
+    // https://api-docs.deepseek.com/guides/thinking_mode
+    { id: "deepseek-flash", name: "deepseek-flash", providerId: "deepseek", reasoningLevels: DEEPSEEK_V4_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    // Legacy vision alias now routes to V4.1 Flash. Keep its explicit-only
+    // routing policy; existing sessions can continue using the old model ID.
+    { id: "deepseek-v4-flash-vision-exp", name: "DeepSeek-V4-Flash-Vision-Exp", providerId: "deepseek", reasoningLevels: DEEPSEEK_V4_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "deepseek-v4-pro", name: "deepseek-v4-pro", providerId: "deepseek", tier: "strong", reasoningLevels: DEEPSEEK_V4_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    // Offline/no-key fallback only: with an API key the registry replaces this
+    // list via fetchGeminiModels (GET /v1beta/models, newest five). Gemini 3
+    // thinking support varies by release: 3.7/3.8 Flash support off/low/medium/high
+    // while older Flash releases expose minimal/low/medium/high.
+    // Gemini 4 Argon: 1,048,576-token window shared by input, thinking and output.
+    // https://ai.google.dev/gemini-api/docs/models/gemini-4-argon
+    { id: "gemini-4-argon", name: "Gemini 4 Argon", providerId: "google", tier: "strong", reasoningLevels: GEMINI_4_LEVELS, defaultReasoningLevel: "medium", contextWindow: 1048576 },
+    // Gemini 4 Carbon (2026-11-10): Google's most capable reasoning model, same
+    // window, levels and medium default as Argon (xhigh verified 400 on 2026-11-12).
+    // Listed after Argon so the provider's default model does not change.
+    // https://ai.google.dev/gemini-api/docs/models/gemini-4-carbon
+    { id: "gemini-4-carbon", name: "Gemini 4 Carbon", providerId: "google", tier: "strong", reasoningLevels: GEMINI_4_LEVELS, defaultReasoningLevel: "medium", contextWindow: 1048576 },
+    // Gemini 4 Hydrogen (2026-11-12): same shared 1,048,576-token window, 65,536
+    // output by default. https://ai.google.dev/gemini-api/docs/models/gemini-4-hydrogen
+    { id: "gemini-4-hydrogen", name: "Gemini 4 Hydrogen", providerId: "google", tier: "balanced", reasoningLevels: GEMINI_4_HYDROGEN_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "gemini-3.8-flash", name: "Gemini 3.8 Flash", providerId: "google", tier: "fast", reasoningLevels: GEMINI_37_38_FLASH_LEVELS, contextWindow: 1048576 },
+    { id: "gemini-3.7-flash", name: "Gemini 3.7 Flash", providerId: "google", tier: "fast", reasoningLevels: GEMINI_37_38_FLASH_LEVELS, contextWindow: 1048576 },
+    { id: "gemini-3.5-flash", name: "Gemini 3.5 Flash", providerId: "google", tier: "fast", reasoningLevels: GEMINI_3_FLASH_LEVELS, contextWindow: 1048576 },
+    { id: "gemini-3.1-pro-preview", name: "Gemini 3.1 Pro", providerId: "google", tier: "strong", reasoningLevels: GEMINI_3_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "gemini-3-flash-preview", name: "Gemini 3 Flash", providerId: "google", tier: "fast", reasoningLevels: GEMINI_3_FLASH_LEVELS, contextWindow: 1048576 },
+    { id: "gemini-2.5-pro", name: "Gemini 2.5 Pro", providerId: "google", tier: "strong", reasoningLevels: GEMINI_25_PRO_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "gemini-2.5-flash", name: "Gemini 2.5 Flash", providerId: "google", tier: "fast", reasoningLevels: GEMINI_25_FLASH_LEVELS, contextWindow: 1048576 },
+    { id: "glm-5.2", name: "GLM-5.2", providerId: "zhipuai", tier: "strong", reasoningLevels: GLM_5_2_LEVELS, contextWindow: 1000000 },
+    { id: "glm-5.1", name: "GLM-5.1", providerId: "zhipuai", tier: "balanced", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 200000 },
+    { id: "glm-4.7", name: "GLM-4.7", providerId: "zhipuai", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 204800 },
+    { id: "glm-4.6", name: "GLM-4.6", providerId: "zhipuai", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 204800 },
+    // GLM-5.3-Flash is a native multimodal model with the same mandatory
+    // low/high/max thinking contract as GLM-5.3. It is fully available on the
+    // Coding Plan and GET /models (verified live 2026-08-26). Official docs list
+    // a 1M context window and position it as the fast/low-cost plan model.
+    { id: "glm-5.3-flash", name: "GLM-5.3-Flash", providerId: "zhipuai-coding-plan", tier: "fast", reasoningLevels: GLM_5_3_LEVELS, defaultReasoningLevel: "max", contextWindow: 1000000 },
+    // Coding Plan exposed GLM-5.3 before GET /models listed it (verified live
+    // 2026-08-14). The standard API was still marked "coming soon" at that time.
+    { id: "glm-5.3", name: "GLM-5.3", providerId: "zhipuai-coding-plan", tier: "strong", reasoningLevels: GLM_5_3_LEVELS, defaultReasoningLevel: "max", contextWindow: 1000000 },
+    { id: "glm-5.2", name: "GLM-5.2", providerId: "zhipuai-coding-plan", tier: "strong", reasoningLevels: GLM_5_2_LEVELS, contextWindow: 1000000 },
+    { id: "glm-5.1", name: "GLM-5.1", providerId: "zhipuai-coding-plan", tier: "balanced", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 200000 },
+    { id: "glm-4.7", name: "GLM-4.7", providerId: "zhipuai-coding-plan", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 204800 },
+    { id: "glm-4.6", name: "GLM-4.6", providerId: "zhipuai-coding-plan", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 204800 },
+    { id: "glm-5.2", name: "GLM-5.2", providerId: "zai", tier: "strong", reasoningLevels: GLM_5_2_LEVELS, contextWindow: 1000000 },
+    { id: "glm-5.1", name: "GLM-5.1", providerId: "zai", tier: "balanced", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 200000 },
+    { id: "glm-4.7", name: "GLM-4.7", providerId: "zai", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 204800 },
+    { id: "glm-4.6", name: "GLM-4.6", providerId: "zai", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 204800 },
+    { id: "glm-5.2", name: "GLM-5.2", providerId: "zai-coding-plan", tier: "strong", reasoningLevels: GLM_5_2_LEVELS, contextWindow: 1000000 },
+    { id: "glm-5-turbo", name: "GLM-5-Turbo", providerId: "zai-coding-plan", tier: "fast", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 200000 },
+    { id: "glm-4.7", name: "GLM-4.7", providerId: "zai-coding-plan", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 204800 },
+    { id: "glm-4.6", name: "GLM-4.6", providerId: "zai-coding-plan", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 200000 },
+    { id: "qwen3.6-plus", name: "Qwen3.6 Plus", providerId: "alibaba", tier: "balanced", reasoningLevels: ["off"], contextWindow: 1048576 },
+    { id: "qwen3.7-max", name: "Qwen3.7 Max", providerId: "alibaba", tier: "strong", reasoningLevels: ["off"], contextWindow: 1048576 },
+    // DashScope Kimi, first-party ids (probed 2026-10-26: kimi-k3 tiers track
+    // upstream, k2.5/k2.6 default to thinking off, K2.7 Code always thinks).
+    { id: "kimi-k3", name: "Kimi K3", providerId: "alibaba", tier: "strong", reasoningLevels: KIMI_K3_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "kimi-k2.7-code", name: "Kimi K2.7 Code", providerId: "alibaba", tier: "strong", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi-k2.6", name: "Kimi K2.6", providerId: "alibaba", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 256000 },
+    { id: "kimi-k2.5", name: "Kimi K2.5", providerId: "alibaba", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 256000 },
+    // Moonshot passthrough ids, per DashScope's Kimi doc: K3 is max-only, K2.7
+    // Code always thinks, K2.6 toggles via enable_thinking.
+    { id: "kimi/kimi-k3.1-pro", name: "Kimi K3.1 Pro", providerId: "alibaba", tier: "strong", reasoningLevels: KIMI_PASSTHROUGH_K3_LEVELS, defaultReasoningLevel: "max", contextWindow: 1048576 },
+    { id: "kimi/kimi-k3.1-pro-highspeed", name: "Kimi K3.1 Pro Highspeed", providerId: "alibaba", tier: "fast", reasoningLevels: KIMI_PASSTHROUGH_K3_LEVELS, defaultReasoningLevel: "max", contextWindow: 1048576 },
+    { id: "kimi/kimi-k3", name: "Kimi K3 (Moonshot)", providerId: "alibaba", tier: "strong", reasoningLevels: KIMI_PASSTHROUGH_K3_LEVELS, defaultReasoningLevel: "max", contextWindow: 1048576 },
+    { id: "kimi/kimi-k2.7-code", name: "Kimi K2.7 Code (Moonshot)", providerId: "alibaba", tier: "strong", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi/kimi-k2.7-code-highspeed", name: "Kimi K2.7 Code Highspeed", providerId: "alibaba", tier: "fast", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi/kimi-k2.6", name: "Kimi K2.6 (Moonshot)", providerId: "alibaba", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 256000 },
+    // Bailian token plan (GET /models on the plan endpoint, probed 2026-08-04).
+    // Image/TTS entries from that list are omitted — this is a text agent.
+    { id: "qwen3.8-max", name: "Qwen3.8 Max", providerId: "bailian-token-plan", tier: "strong", reasoningLevels: BAILIAN_FULL_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "qwen3.8-max-preview", name: "Qwen3.8 Max Preview", providerId: "bailian-token-plan", tier: "strong", reasoningLevels: BAILIAN_FULL_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "qwen3.7-max", name: "Qwen3.7 Max", providerId: "bailian-token-plan", tier: "strong", reasoningLevels: BAILIAN_NO_MAX_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "qwen3.7-plus", name: "Qwen3.7 Plus", providerId: "bailian-token-plan", tier: "balanced", reasoningLevels: BAILIAN_NO_MAX_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "qwen3.6-flash", name: "Qwen3.6 Flash", providerId: "bailian-token-plan", tier: "fast", reasoningLevels: BAILIAN_NO_MAX_LEVELS, defaultReasoningLevel: "medium", contextWindow: 1048576 },
+    { id: "glm-5.2", name: "GLM-5.2", providerId: "bailian-token-plan", tier: "strong", reasoningLevels: BAILIAN_FULL_LEVELS, defaultReasoningLevel: "high", contextWindow: 1000000 },
+    { id: "deepseek-v4-pro", name: "DeepSeek V4 Pro", providerId: "bailian-token-plan", tier: "strong", reasoningLevels: BAILIAN_DEEPSEEK_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "deepseek-v4-flash-0731", name: "DeepSeek V4 Flash 0731", providerId: "bailian-token-plan", tier: "fast", reasoningLevels: BAILIAN_DEEPSEEK_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "doubao-seed-2-1-pro-260628", name: "Doubao Seed 2.1 Pro", providerId: "doubao", reasoningLevels: DOUBAO_SEED_REASONING_LEVELS, defaultReasoningLevel: "high" },
+    { id: "MiniMax-M3", name: "MiniMax M3", providerId: "minimax", tier: "strong", reasoningLevels: MINIMAX_M3_REASONING_LEVELS, contextWindow: 1000000 },
+    { id: "MiniMax-M2.7", name: "MiniMax M2.7", providerId: "minimax", tier: "balanced", reasoningLevels: MINIMAX_REASONING_LEVELS, contextWindow: 204800 },
+    { id: "MiniMax-M2.7-highspeed", name: "MiniMax M2.7 Highspeed", providerId: "minimax", tier: "fast", reasoningLevels: MINIMAX_REASONING_LEVELS, contextWindow: 204800 },
+    { id: "MiniMax-M3", name: "MiniMax M3", providerId: "minimax-anthropic", tier: "strong", reasoningLevels: MINIMAX_M3_REASONING_LEVELS, contextWindow: 1000000 },
+    { id: "MiniMax-M2.7", name: "MiniMax M2.7", providerId: "minimax-anthropic", tier: "balanced", reasoningLevels: MINIMAX_REASONING_LEVELS, contextWindow: 204800 },
+    { id: "MiniMax-M2.7-highspeed", name: "MiniMax M2.7 Highspeed", providerId: "minimax-anthropic", tier: "fast", reasoningLevels: MINIMAX_REASONING_LEVELS, contextWindow: 204800 },
+    // Token Plan's text models; V2.5 remains available until 2026-10-21.
+    // https://mimo.mi.com/docs/zh-CN/tokenplan/integration/claudecode
+    { id: "mimo-v2.6-pro", name: "MiMo V2.6 Pro", providerId: "mimo-token-plan", tier: "strong", reasoningLevels: TOGGLE_THINKING_LEVELS, defaultReasoningLevel: "medium", contextWindow: 1000000 },
+    { id: "mimo-v2.6-flash", name: "MiMo V2.6 Flash", providerId: "mimo-token-plan", tier: "fast", reasoningLevels: TOGGLE_THINKING_LEVELS, defaultReasoningLevel: "medium", contextWindow: 1000000 },
+    { id: "mimo-v2.5-pro", name: "MiMo V2.5 Pro", providerId: "mimo-token-plan", reasoningLevels: TOGGLE_THINKING_LEVELS, defaultReasoningLevel: "medium", contextWindow: 1000000 },
+    { id: "mimo-v2.5", name: "MiMo V2.5", providerId: "mimo-token-plan", reasoningLevels: TOGGLE_THINKING_LEVELS, defaultReasoningLevel: "medium", contextWindow: 1000000 },
+    { id: "step-3.7-flash", name: "Step 3.7 Flash", providerId: "stepfun", reasoningLevels: STEPFUN_REASONING_LEVELS, contextWindow: 256000 },
+    { id: "step-3.5-flash-2603", name: "Step 3.5 Flash 2603", providerId: "stepfun", reasoningLevels: STEPFUN_REASONING_LEVELS },
+    { id: "step-3.5-flash", name: "Step 3.5 Flash", providerId: "stepfun", reasoningLevels: STEPFUN_REASONING_LEVELS },
+    { id: "step-router-v1", name: "Step Router V1", providerId: "stepfun", reasoningLevels: STEPFUN_REASONING_LEVELS },
+    // Use server-default reasoning until this model's effort controls are documented.
+    { id: "water18-0910", name: "water18-0910", providerId: "stepfun-api", reasoningLevels: [] },
+    { id: "kimi-k2.7-code", name: "Kimi K2.7 Code", providerId: "moonshot-cn", tier: "strong", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi-k2.7-code-highspeed", name: "Kimi K2.7 Code Highspeed", providerId: "moonshot-cn", tier: "fast", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi-k3", name: "Kimi K3", providerId: "moonshot-cn", tier: "strong", reasoningLevels: KIMI_K3_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "kimi-k3.1-pro", name: "Kimi K3.1 Pro", providerId: "moonshot-cn", tier: "strong", reasoningLevels: KIMI_K3_EFFORT_ONLY_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "kimi-k3.1-pro-highspeed", name: "Kimi K3.1 Pro Highspeed", providerId: "moonshot-cn", tier: "fast", reasoningLevels: KIMI_K3_EFFORT_ONLY_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "kimi-k2.6", name: "Kimi K2.6", providerId: "moonshot-cn", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 256000 },
+    { id: "kimi-k2.5", name: "Kimi K2.5", providerId: "moonshot-cn", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 256000 },
+    { id: "kimi-k2.7-code", name: "Kimi K2.7 Code", providerId: "moonshot-intl", tier: "strong", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi-k2.7-code-highspeed", name: "Kimi K2.7 Code Highspeed", providerId: "moonshot-intl", tier: "fast", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi-k3", name: "Kimi K3", providerId: "moonshot-intl", tier: "strong", reasoningLevels: KIMI_K3_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "kimi-k3.1-pro", name: "Kimi K3.1 Pro", providerId: "moonshot-intl", tier: "strong", reasoningLevels: KIMI_K3_EFFORT_ONLY_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "kimi-k3.1-pro-highspeed", name: "Kimi K3.1 Pro Highspeed", providerId: "moonshot-intl", tier: "fast", reasoningLevels: KIMI_K3_EFFORT_ONLY_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "kimi-k2.6", name: "Kimi K2.6", providerId: "moonshot-intl", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 256000 },
+    { id: "kimi-k2.5", name: "Kimi K2.5", providerId: "moonshot-intl", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 256000 },
+    // Kimi for Coding serves plan-slot ids (verified via GET /coding/v1/models):
+    // k3, k3-256k, k3.1-pro, k3.1-pro-highspeed, kimi-for-coding,
+    // kimi-for-coding-highspeed.
+    { id: "k3", name: "Kimi K3", providerId: "kimi-for-coding", tier: "strong", reasoningLevels: KIMI_K3_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "k3-256k", name: "Kimi K3 256K", providerId: "kimi-for-coding", tier: "strong", reasoningLevels: KIMI_K3_EFFORT_ONLY_LEVELS, defaultReasoningLevel: "high", contextWindow: 262144 },
+    { id: "k3.1-pro", name: "Kimi K3.1 Pro", providerId: "kimi-for-coding", tier: "strong", reasoningLevels: KIMI_K3_EFFORT_ONLY_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "k3.1-pro-highspeed", name: "Kimi K3.1 Pro Highspeed", providerId: "kimi-for-coding", tier: "fast", reasoningLevels: KIMI_K3_EFFORT_ONLY_LEVELS, defaultReasoningLevel: "high", contextWindow: 1048576 },
+    { id: "kimi-for-coding", name: "Kimi for Coding", providerId: "kimi-for-coding", tier: "strong", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi-for-coding-highspeed", name: "Kimi for Coding Highspeed", providerId: "kimi-for-coding", tier: "fast", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi-k2.7-code", name: "Kimi K2.7 Code", providerId: "kimi-for-coding", tier: "strong", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi-k2.7-code-highspeed", name: "Kimi K2.7 Code Highspeed", providerId: "kimi-for-coding", tier: "fast", reasoningLevels: KIMI_THINKING_ONLY_LEVELS, contextWindow: 262144 },
+    { id: "kimi-k2.6", name: "Kimi K2.6", providerId: "kimi-for-coding", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 256000 },
+    { id: "kimi-k2.5", name: "Kimi K2.5", providerId: "kimi-for-coding", reasoningLevels: TOGGLE_THINKING_LEVELS, contextWindow: 256000 },
+    { id: "llama-3.3-70b-versatile", name: "llama-3.3-70b-versatile", providerId: "groq", reasoningLevels: ["off"], contextWindow: 131072 },
+    { id: "mixtral-8x7b-32768", name: "mixtral-8x7b-32768", providerId: "groq", reasoningLevels: ["off"], contextWindow: 32768 },
+    { id: "gemma-2-9b-it", name: "gemma-2-9b-it", providerId: "groq", reasoningLevels: ["off"], contextWindow: 32768 },
+    { id: "meta-llama/Llama-3.3-70B-Instruct-Turbo", name: "meta-llama/Llama-3.3-70B-Instruct-Turbo", providerId: "together", reasoningLevels: ["off"], contextWindow: 131072 },
+    { id: "Qwen/Qwen2.5-72B-Instruct", name: "Qwen/Qwen2.5-72B-Instruct", providerId: "together", reasoningLevels: ["off"], contextWindow: 32768 },
+    { id: "accounts/fireworks/models/kimi-k2p6", name: "Kimi-K2.6", providerId: "fireworks", reasoningLevels: ["off"], contextWindow: 256000 },
+    { id: "llama3.1", name: "llama3.1", providerId: "local", reasoningLevels: ["off"], contextWindow: 32768 },
+    { id: "qwen2.5", name: "qwen2.5", providerId: "local", reasoningLevels: ["off"], contextWindow: 32768 },
+    { id: "deepseek-coder-v2", name: "deepseek-coder-v2", providerId: "local", reasoningLevels: ["off"], contextWindow: 32768 },
+];
+export function listBuiltinModels(providerId) {
+    return BUILTIN_MODELS.filter((model) => model.providerId === providerId);
+}
+// Runtime overlay populated from provider-side discovery (e.g. ChatGPT codex /models).
+// Looked up before the static catalog so newly-released models work without a code change.
+const dynamicOverlay = new Map();
+function overlayKey(providerId, modelId) {
+    return `${providerId}:${modelId}`;
+}
+export function registerDynamicModelMetadata(model) {
+    dynamicOverlay.set(overlayKey(model.providerId, model.id), model);
+}
+export function clearDynamicModelMetadata(providerId) {
+    const providerIds = providerId === "openai"
+        ? ["openai", "openai-codex"]
+        : [providerId];
+    for (const key of dynamicOverlay.keys()) {
+        if (providerIds.some((id) => key.startsWith(`${id}:`)))
+            dynamicOverlay.delete(key);
+    }
+}
+export function replaceDynamicModelMetadata(providerId, models) {
+    clearDynamicModelMetadata(providerId);
+    const acceptedProviderIds = providerId === "openai"
+        ? new Set(["openai", "openai-codex"])
+        : new Set([providerId]);
+    for (const model of models) {
+        if (acceptedProviderIds.has(model.providerId))
+            registerDynamicModelMetadata(model);
+    }
+}
+/**
+ * All dynamic-overlay entries for a provider (openai includes its
+ * openai-codex catalog alias). The overlay was previously visible only via
+ * getBuiltinModel point lookups; the routing snapshot builder needs the list.
+ */
+export function listDynamicModelMetadata(providerId) {
+    const providerIds = providerId === "openai" ? ["openai", "openai-codex"] : [providerId];
+    const out = [];
+    for (const [key, model] of dynamicOverlay) {
+        if (providerIds.some((id) => key.startsWith(`${id}:`)))
+            out.push(model);
+    }
+    return out;
+}
+export function getBuiltinModel(providerId, modelId) {
+    const overlayHit = dynamicOverlay.get(overlayKey(providerId, modelId))
+        || (providerId === "openai" ? dynamicOverlay.get(overlayKey("openai-codex", modelId)) : undefined);
+    if (overlayHit)
+        return overlayHit;
+    const staticHit = BUILTIN_MODELS.find((model) => model.providerId === providerId && model.id === modelId)
+        || (providerId === "openai"
+            ? BUILTIN_MODELS.find((model) => model.providerId === "openai-codex" && model.id === modelId)
+            : undefined);
+    if (staticHit)
+        return staticHit;
+    // Dynamically-discovered grok models aren't in the static catalog until
+    // remote discovery runs. Synthesize their metadata from the id so reasoning
+    // levels, default effort, and context window resolve correctly on a fresh
+    // start or resumed session (before the first /model discovery).
+    if (providerId === "grok") {
+        const inferred = inferGrokModelMetadata(modelId);
+        return {
+            id: modelId,
+            name: modelId,
+            providerId: "grok",
+            reasoningLevels: inferred.levels,
+            defaultReasoningLevel: inferred.defaultLevel,
+            contextWindow: inferred.contextWindow,
+        };
+    }
+    return undefined;
+}
+export function getModelDefaultReasoningLevel(providerId, modelId) {
+    return getBuiltinModel(providerId, modelId)?.defaultReasoningLevel;
+}
+export function getBuiltinProvider(providerId) {
+    return BUILTIN_PROVIDERS.find((provider) => provider.id === providerId);
+}
+/**
+ * Infer capability metadata for a Grok model id the /models endpoint returned
+ * without metadata. Flagship reasoning models (grok-N.M) expose low/medium/high
+ * with thinking always on and a 500K context window — matching the curated
+ * grok-4.5 entry — while composer/fast/mini/flash variants have no effort
+ * control and a 200K window. Pure (id-only) so it works even before remote
+ * model discovery runs.
+ */
+export function inferGrokModelMetadata(modelId) {
+    const id = modelId.toLowerCase();
+    if (/(composer|fast|mini|flash)/.test(id)) {
+        return { levels: ["off"], contextWindow: 200000 };
+    }
+    return { levels: ["low", "medium", "high"], defaultLevel: "high", contextWindow: 500000 };
+}
+export function getModelContextWindow(providerId, modelId) {
+    const builtin = getBuiltinModel(providerId, modelId);
+    if (builtin?.contextWindow !== undefined)
+        return builtin.contextWindow;
+    // Dynamically-discovered grok models aren't in the static catalog until
+    // discovery runs; infer the window from the id so context usage can render a
+    // percentage even before discovery (e.g. on a resumed session right after
+    // startup).
+    if (providerId === "grok")
+        return inferGrokModelMetadata(modelId).contextWindow;
+    return undefined;
+}
+export function getToolOutputTokenLimit(providerId, modelId) {
+    return getBuiltinModel(providerId, modelId)?.toolOutputTokenLimit;
+}

package/dist/model-config.d.ts

@@ -0,0 +1,40 @@
+/**
+ * models.json loader.
+ *
+ * Users can define providers, API keys, base URLs, and custom models
+ * in ~/.bubble/models.json.
+ */
+import type { ModelTier, ProviderProtocol } from "./model-catalog.js";
+import type { ModelInfo } from "./provider-registry.js";
+export interface ProviderModelConfig {
+    baseURL?: string;
+    apiKey?: string;
+    protocol?: ProviderProtocol;
+    /** Extra request headers (e.g. a User-Agent a coding-plan endpoint gates on). */
+    headers?: Record<string, string>;
+    models?: Array<{
+        id: string;
+        name?: string;
+        tier?: ModelTier;
+        routingPriority?: number;
+    }>;
+}
+export interface ModelsConfig {
+    providers: Record<string, ProviderModelConfig>;
+}
+export declare class ModelConfig {
+    private readonly modelsPath;
+    private data?;
+    private loadError?;
+    constructor();
+    private load;
+    getLoadError(): string | undefined;
+    getPath(): string;
+    getProviderConfig(providerId: string): ProviderModelConfig | undefined;
+    getAllProviders(): Record<string, ProviderModelConfig>;
+    hasProvider(providerId: string): boolean;
+    getCustomModels(providerId: string): ModelInfo[];
+    getApiKey(providerId: string): string | undefined;
+    getBaseURL(providerId: string): string | undefined;
+    getProtocol(providerId: string): ProviderProtocol | undefined;
+}

package/dist/model-config.js

@@ -0,0 +1,74 @@
+/**
+ * models.json loader.
+ *
+ * Users can define providers, API keys, base URLs, and custom models
+ * in ~/.bubble/models.json.
+ */
+import { existsSync, readFileSync } from "node:fs";
+import { join } from "node:path";
+import { getBubbleHome } from "./bubble-home.js";
+import { filterProviderModels } from "./provider-model-policy.js";
+const MODEL_TIERS = ["fast", "balanced", "strong"];
+function sanitizeTier(value) {
+    return typeof value === "string" && MODEL_TIERS.includes(value)
+        ? value
+        : undefined;
+}
+function sanitizeRoutingPriority(value) {
+    return typeof value === "number" && Number.isFinite(value) ? value : undefined;
+}
+export class ModelConfig {
+    modelsPath = join(getBubbleHome(), "models.json");
+    data;
+    loadError;
+    constructor() {
+        this.load();
+    }
+    load() {
+        if (!existsSync(this.modelsPath))
+            return;
+        try {
+            const raw = readFileSync(this.modelsPath, "utf-8");
+            this.data = JSON.parse(raw);
+        }
+        catch (err) {
+            this.loadError = err.message;
+        }
+    }
+    getLoadError() {
+        return this.loadError;
+    }
+    getPath() {
+        return this.modelsPath;
+    }
+    getProviderConfig(providerId) {
+        return this.data?.providers?.[providerId];
+    }
+    getAllProviders() {
+        return this.data?.providers ?? {};
+    }
+    hasProvider(providerId) {
+        return !!this.data?.providers?.[providerId];
+    }
+    getCustomModels(providerId) {
+        const cfg = this.data?.providers?.[providerId];
+        if (!cfg?.models)
+            return [];
+        return filterProviderModels(providerId, cfg.models.map((m) => ({
+            id: m.id,
+            name: m.name || m.id,
+            providerId,
+            tier: sanitizeTier(m.tier),
+            routingPriority: sanitizeRoutingPriority(m.routingPriority),
+        })));
+    }
+    getApiKey(providerId) {
+        return this.data?.providers?.[providerId]?.apiKey;
+    }
+    getBaseURL(providerId) {
+        return this.data?.providers?.[providerId]?.baseURL;
+    }
+    getProtocol(providerId) {
+        return this.data?.providers?.[providerId]?.protocol;
+    }
+}

package/dist/model-data.generated.d.ts

@@ -0,0 +1,16 @@
+export interface GeneratedModelData {
+    providerId: string;
+    modelId: string;
+    name: string;
+    /** USD per million tokens. */
+    inputPerMillion: number;
+    /** USD per million tokens. */
+    outputPerMillion: number;
+    /** USD per million tokens; absent when the provider has no cache discount. */
+    cacheReadPerMillion?: number;
+    /** USD per million tokens; absent when cache writes bill at the input rate. */
+    cacheWritePerMillion?: number;
+    contextWindow?: number;
+    maxOutputTokens?: number;
+}
+export declare const GENERATED_MODEL_DATA: GeneratedModelData[];