@bastani/pi-ai 0.9.19 → 0.9.20-alpha.10

package/dist/providers/data/xai.json

@@ -1 +1 @@
-{"openai-responses":{"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","api":"openai-responses","provider":"xai","baseUrl":"https://api.x.ai/v1","compat":{"supportsLongCacheRetention":false},"reasoning":true,"input":["text","image"],"cost":{"input":1.25,"output":2.5,"cacheRead":0.2,"cacheWrite":0},"contextWindow":1000000,"maxTokens":30000,"thinkingLevelMap":{"off":"none","minimal":null,"low":"low","medium":"medium","high":"high","xhigh":null,"max":null}},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","api":"openai-responses","provider":"xai","baseUrl":"https://api.x.ai/v1","compat":{"supportsLongCacheRetention":false},"reasoning":true,"input":["text","image"],"cost":{"input":2,"output":6,"cacheRead":0.3,"cacheWrite":0},"contextWindow":500000,"maxTokens":500000,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":"medium","high":"high","xhigh":null,"max":null}},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","api":"openai-responses","provider":"xai","baseUrl":"https://api.x.ai/v1","compat":{"supportsLongCacheRetention":false},"reasoning":true,"input":["text","image"],"cost":{"input":2,"output":6,"cacheRead":0.5,"cacheWrite":0},"contextWindow":500000,"maxTokens":500000,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":"medium","high":"high","xhigh":"xhigh","max":null}}}}
+{"openai-responses":{"chat:grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","api":"openai-responses","provider":"xai","baseUrl":"https://api.x.ai/v1","compat":{"supportsLongCacheRetention":false},"reasoning":true,"input":["text","image"],"cost":{"input":1.25,"output":2.5,"cacheRead":0.2,"cacheWrite":0,"tiers":[{"inputTokensAbove":200000,"input":2.5,"output":5,"cacheRead":0.4,"cacheWrite":0}]},"contextWindow":1000000,"maxTokens":30000,"thinkingLevelMap":{"off":"none","minimal":null,"low":"low","medium":"medium","high":"high","xhigh":null,"max":null},"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","api":"openai-responses","provider":"xai","baseUrl":"https://api.x.ai/v1","compat":{"supportsLongCacheRetention":false},"reasoning":true,"input":["text","image"],"cost":{"input":2,"output":6,"cacheRead":0.3,"cacheWrite":0,"tiers":[{"inputTokensAbove":200000,"input":4,"output":12,"cacheRead":0.6,"cacheWrite":0}]},"contextWindow":500000,"maxTokens":500000,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":"medium","high":"high","xhigh":null,"max":null},"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","api":"openai-responses","provider":"xai","baseUrl":"https://api.x.ai/v1","compat":{"supportsLongCacheRetention":false},"reasoning":true,"input":["text","image"],"cost":{"input":2,"output":6,"cacheRead":0.5,"cacheWrite":0,"tiers":[{"inputTokensAbove":200000,"input":4,"output":12,"cacheRead":1,"cacheWrite":0}]},"contextWindow":500000,"maxTokens":500000,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":"medium","high":"high","xhigh":"xhigh","max":null},"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","api":"openai-responses","provider":"xai","baseUrl":"https://api.x.ai/v1","compat":{"supportsLongCacheRetention":false},"reasoning":true,"input":["text","image"],"cost":{"input":2,"output":6,"cacheRead":0.5,"cacheWrite":0,"tiers":[{"inputTokensAbove":200000,"input":4,"output":12,"cacheRead":1,"cacheWrite":0}]},"contextWindow":500000,"maxTokens":500000,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":"medium","high":"high","xhigh":"xhigh","max":null},"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"}}}

package/dist/providers/data/xiaomi-token-plan-ams.json

@@ -1 +1 @@
-{"openai-completions":{"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","api":"openai-completions","provider":"xiaomi-token-plan-ams","baseUrl":"https://token-plan-ams.xiaomimimo.com/v1","compat":{"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","api":"openai-completions","provider":"xiaomi-token-plan-ams","baseUrl":"https://token-plan-ams.xiaomimimo.com/v1","compat":{"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072}}}
+{"openai-completions":{"chat:mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","api":"openai-completions","provider":"xiaomi-token-plan-ams","baseUrl":"https://token-plan-ams.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","api":"openai-completions","provider":"xiaomi-token-plan-ams","baseUrl":"https://token-plan-ams.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"type":"chat"},"chat:mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","api":"openai-completions","provider":"xiaomi-token-plan-ams","baseUrl":"https://token-plan-ams.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","api":"openai-completions","provider":"xiaomi-token-plan-ams","baseUrl":"https://token-plan-ams.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"}}}

package/dist/providers/data/xiaomi-token-plan-cn.json

@@ -1 +1 @@
-{"openai-completions":{"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","api":"openai-completions","provider":"xiaomi-token-plan-cn","baseUrl":"https://token-plan-cn.xiaomimimo.com/v1","compat":{"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","api":"openai-completions","provider":"xiaomi-token-plan-cn","baseUrl":"https://token-plan-cn.xiaomimimo.com/v1","compat":{"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072}}}
+{"openai-completions":{"chat:mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","api":"openai-completions","provider":"xiaomi-token-plan-cn","baseUrl":"https://token-plan-cn.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","api":"openai-completions","provider":"xiaomi-token-plan-cn","baseUrl":"https://token-plan-cn.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"type":"chat"},"chat:mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","api":"openai-completions","provider":"xiaomi-token-plan-cn","baseUrl":"https://token-plan-cn.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","api":"openai-completions","provider":"xiaomi-token-plan-cn","baseUrl":"https://token-plan-cn.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"}}}

package/dist/providers/data/xiaomi-token-plan-sgp.json

@@ -1 +1 @@
-{"openai-completions":{"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","api":"openai-completions","provider":"xiaomi-token-plan-sgp","baseUrl":"https://token-plan-sgp.xiaomimimo.com/v1","compat":{"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","api":"openai-completions","provider":"xiaomi-token-plan-sgp","baseUrl":"https://token-plan-sgp.xiaomimimo.com/v1","compat":{"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072}}}
+{"openai-completions":{"chat:mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","api":"openai-completions","provider":"xiaomi-token-plan-sgp","baseUrl":"https://token-plan-sgp.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","api":"openai-completions","provider":"xiaomi-token-plan-sgp","baseUrl":"https://token-plan-sgp.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"type":"chat"},"chat:mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","api":"openai-completions","provider":"xiaomi-token-plan-sgp","baseUrl":"https://token-plan-sgp.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","api":"openai-completions","provider":"xiaomi-token-plan-sgp","baseUrl":"https://token-plan-sgp.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"}}}

package/dist/providers/data/xiaomi.json

@@ -1 +1 @@
-{"openai-completions":{"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","api":"openai-completions","provider":"xiaomi","baseUrl":"https://api.xiaomimimo.com/v1","compat":{"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0.14,"output":0.28,"cacheRead":0.0028,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","api":"openai-completions","provider":"xiaomi","baseUrl":"https://api.xiaomimimo.com/v1","compat":{"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":0.435,"output":0.87,"cacheRead":0.0036,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072},"mimo-v2.5-pro-ultraspeed":{"id":"mimo-v2.5-pro-ultraspeed","name":"MiMo-V2.5-Pro-UltraSpeed","api":"openai-completions","provider":"xiaomi","baseUrl":"https://api.xiaomimimo.com/v1","compat":{"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":1.305,"output":2.61,"cacheRead":0.0108,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072}}}
+{"openai-completions":{"chat:mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","api":"openai-completions","provider":"xiaomi","baseUrl":"https://api.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0.14,"output":0.28,"cacheRead":0.0028,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","api":"openai-completions","provider":"xiaomi","baseUrl":"https://api.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":0.435,"output":0.87,"cacheRead":0.0036,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"type":"chat"},"chat:mimo-v2.5-pro-ultraspeed":{"id":"mimo-v2.5-pro-ultraspeed","name":"MiMo-V2.5-Pro-UltraSpeed","api":"openai-completions","provider":"xiaomi","baseUrl":"https://api.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text"],"cost":{"input":1.305,"output":2.61,"cacheRead":0.0108,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"type":"chat"},"chat:mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","api":"openai-completions","provider":"xiaomi","baseUrl":"https://api.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0.14,"output":0.28,"cacheRead":0.0028,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","api":"openai-completions","provider":"xiaomi","baseUrl":"https://api.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":0.435,"output":0.87,"cacheRead":0.0036,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:mimo-v2.6-pro-ultraspeed":{"id":"mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","api":"openai-completions","provider":"xiaomi","baseUrl":"https://api.xiaomimimo.com/v1","compat":{"supportsStrictMode":true,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},"reasoning":true,"input":["text","image"],"cost":{"input":4.35,"output":8.7,"cacheRead":0.036,"cacheWrite":0},"contextWindow":1048576,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"}}}

package/dist/providers/data/zai-coding-cn.json

@@ -1 +1 @@
-{"openai-completions":{"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"input":["text","image"],"cost":{"input":0.3,"output":0.9,"cacheRead":0,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":128000,"maxTokens":32768},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"input":["text"],"cost":{"input":0.6,"output":2.2,"cacheRead":0.11,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":204800,"maxTokens":131072},"glm-5-turbo":{"id":"glm-5-turbo","name":"GLM-5-Turbo","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"input":["text"],"cost":{"input":1.2,"output":4,"cacheRead":0.24,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":200000,"maxTokens":131072},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"input":["text"],"cost":{"input":1.4,"output":4.4,"cacheRead":0.26,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":200000,"maxTokens":131072},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":"none","minimal":null,"low":null,"medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":1.4,"output":4.4,"cacheRead":0.26,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072},"glm-5.2-highspeed":{"id":"glm-5.2-highspeed","name":"GLM-5.2 Highspeed","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":"none","minimal":null,"low":null,"medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":1.4,"output":4.4,"cacheRead":0.26,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text","image"],"cost":{"input":0.075,"output":0.25,"cacheRead":0.015,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072},"glm-5.3-highspeed":{"id":"glm-5.3-highspeed","name":"GLM-5.3 Highspeed","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM-5V-Turbo","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"input":["text","image"],"cost":{"input":1.2,"output":4,"cacheRead":0.24,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":200000,"maxTokens":131072}}}
+{"openai-completions":{"chat:glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"input":["text","image"],"cost":{"input":0.3,"output":0.9,"cacheRead":0,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":128000,"maxTokens":32768,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":1.4,"output":4.4,"cacheRead":0.26,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072,"type":"chat"},"chat:glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text","image"],"cost":{"input":0.15,"output":0.5,"cacheRead":0.03,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:glm-5.3-highspeed":{"id":"glm-5.3-highspeed","name":"GLM-5.3 Highspeed","api":"openai-completions","provider":"zai-coding-cn","baseUrl":"https://open.bigmodel.cn/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072,"type":"chat"}}}

package/dist/providers/data/zai.json

@@ -1 +1 @@
-{"openai-completions":{"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"input":["text"],"cost":{"input":0.6,"output":2.2,"cacheRead":0.11,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":204800,"maxTokens":131072},"glm-5-turbo":{"id":"glm-5-turbo","name":"GLM-5-Turbo","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"input":["text"],"cost":{"input":1.2,"output":4,"cacheRead":0.24,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":200000,"maxTokens":131072},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":"none","minimal":null,"low":null,"medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":1.4,"output":4.4,"cacheRead":0.26,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072},"glm-5.2-highspeed":{"id":"glm-5.2-highspeed","name":"GLM-5.2 Highspeed","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":"none","minimal":null,"low":null,"medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":1.4,"output":4.4,"cacheRead":0.26,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text","image"],"cost":{"input":0.075,"output":0.25,"cacheRead":0.015,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072},"glm-5.3-highspeed":{"id":"glm-5.3-highspeed","name":"GLM-5.3 Highspeed","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072}}}
+{"openai-completions":{"chat:glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"input":["text"],"cost":{"input":0.6,"output":2.2,"cacheRead":0.11,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":204800,"maxTokens":131072,"type":"chat"},"chat:glm-5-turbo":{"id":"glm-5-turbo","name":"GLM-5-Turbo","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"input":["text"],"cost":{"input":1.2,"output":4,"cacheRead":0.24,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":200000,"maxTokens":131072,"type":"chat"},"chat:glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":"none","minimal":null,"low":null,"medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":1.4,"output":4.4,"cacheRead":0.26,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072,"type":"chat"},"chat:glm-5.2-highspeed":{"id":"glm-5.2-highspeed","name":"GLM-5.2 Highspeed","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":"none","minimal":null,"low":null,"medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072,"type":"chat"},"chat:glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":1.4,"output":4.4,"cacheRead":0.26,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072,"type":"chat"},"chat:glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text","image"],"cost":{"input":0.15,"output":0.5,"cacheRead":0.03,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072,"inputLimits":{"images":{"resize":{"maxWidth":2000,"maxHeight":2000,"maxBytes":4718592,"jpegQuality":80}}},"type":"chat"},"chat:glm-5.3-highspeed":{"id":"glm-5.3-highspeed","name":"GLM-5.3 Highspeed","api":"openai-completions","provider":"zai","baseUrl":"https://api.z.ai/api/coding/paas/v4","reasoning":true,"thinkingLevelMap":{"off":null,"minimal":null,"low":"low","medium":null,"high":"high","xhigh":null,"max":"max"},"input":["text"],"cost":{"input":0,"output":0,"cacheRead":0,"cacheWrite":0},"compat":{"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"zai","supportsStrictMode":true,"zaiToolStream":true},"contextWindow":1000000,"maxTokens":131072,"type":"chat"}}}

package/dist/providers/deepseek.models.d.ts

@@ -1,4 +1,5 @@
-import values from "./data/deepseek.json";
-import { type ModelCatalog } from "../model-catalog.ts";
-export declare const DEEPSEEK_MODELS: ModelCatalog<typeof values, "deepseek">;
+import { type ChatModelCatalog, type ImageModelCatalog, type ClassifierModelCatalog } from "../model-catalog.ts";
+export declare const DEEPSEEK_MODELS: ChatModelCatalog<typeof import("./data/deepseek.json", { with: { type: "json" } }), "deepseek">;
+export declare const DEEPSEEK_IMAGE_MODELS: ImageModelCatalog<typeof import("./data/deepseek.json", { with: { type: "json" } }), "deepseek">;
+export declare const DEEPSEEK_CLASSIFIER_MODELS: ClassifierModelCatalog<typeof import("./data/deepseek.json", { with: { type: "json" } }), "deepseek">;
 //# sourceMappingURL=deepseek.models.d.ts.map

package/dist/providers/deepseek.models.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"deepseek.models.d.ts","sourceRoot":"","sources":["../../src/providers/deepseek.models.ts"],"names":[],"mappings":"AAGA,OAAO,MAAM,MAAM,sBAAsB,CAAuB;AAChE,OAAO,EAAuB,KAAK,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAE7E,eAAO,MAAM,eAAe,EAAE,YAAY,CAAC,OAAO,MAAM,EAAE,UAAU,CAC5B,CAAC","sourcesContent":["// This file is auto-generated by scripts/generate-models.ts\n// Do not edit manually - run 'npm run generate-models' to update\n\nimport values from \"./data/deepseek.json\" with { type: \"json\" };\nimport { flattenModelCatalog, type ModelCatalog } from \"../model-catalog.ts\";\n\nexport const DEEPSEEK_MODELS: ModelCatalog<typeof values, \"deepseek\"> =\n\tflattenModelCatalog(\"deepseek\", values);\n"]}
+{"version":3,"file":"deepseek.models.d.ts","sourceRoot":"","sources":["../../src/providers/deepseek.models.ts"],"names":[],"mappings":"AAIA,OAAO,EAAoF,KAAK,gBAAgB,EAAE,KAAK,iBAAiB,EAAE,KAAK,sBAAsB,EAAE,MAAM,qBAAqB,CAAC;AAEnM,eAAO,MAAM,eAAe,EAAE,gBAAgB,CAAC,cAAc,sBAAsB,EAAU,UAAE,IAAI,EAAE,MAAM,IAAE,CAAG,EAAE,UAAU,CAChF,CAAC;AAE7C,eAAO,MAAM,qBAAqB,EAAE,iBAAiB,CAAC,cAAc,sBAAsB,EAAU,UAAE,IAAI,EAAE,MAAM,IAAE,CAAG,EAAE,UAAU,CACtF,CAAC;AAE9C,eAAO,MAAM,0BAA0B,EAAE,sBAAsB,CAAC,cAAc,sBAAsB,EAAU,UAAE,IAAI,EAAE,MAAM,IAAE,CAAG,EAAE,UAAU,CAC3F,CAAC","sourcesContent":["// This file is auto-generated by scripts/generate-models.ts\n// Do not edit manually - run 'npm run generate-models' to update\n\nimport values from \"./data/deepseek.json\" with { type: \"json\" };\nimport { flattenChatModelCatalog, flattenImageModelCatalog, flattenClassifierModelCatalog, type ChatModelCatalog, type ImageModelCatalog, type ClassifierModelCatalog } from \"../model-catalog.ts\";\n\nexport const DEEPSEEK_MODELS: ChatModelCatalog<typeof import(\"./data/deepseek.json\", { with: { type: \"json\" } }), \"deepseek\"> =\n\tflattenChatModelCatalog(\"deepseek\", values);\n\nexport const DEEPSEEK_IMAGE_MODELS: ImageModelCatalog<typeof import(\"./data/deepseek.json\", { with: { type: \"json\" } }), \"deepseek\"> =\n\tflattenImageModelCatalog(\"deepseek\", values);\n\nexport const DEEPSEEK_CLASSIFIER_MODELS: ClassifierModelCatalog<typeof import(\"./data/deepseek.json\", { with: { type: \"json\" } }), \"deepseek\"> =\n\tflattenClassifierModelCatalog(\"deepseek\", values);\n"]}

package/dist/providers/deepseek.models.js

@@ -1,6 +1,8 @@
 // This file is auto-generated by scripts/generate-models.ts
 // Do not edit manually - run 'npm run generate-models' to update
 import values from "./data/deepseek.json" with { type: "json" };
-import { flattenModelCatalog } from "../model-catalog.js";
-export const DEEPSEEK_MODELS = flattenModelCatalog("deepseek", values);
+import { flattenChatModelCatalog, flattenImageModelCatalog, flattenClassifierModelCatalog } from "../model-catalog.js";
+export const DEEPSEEK_MODELS = flattenChatModelCatalog("deepseek", values);
+export const DEEPSEEK_IMAGE_MODELS = flattenImageModelCatalog("deepseek", values);
+export const DEEPSEEK_CLASSIFIER_MODELS = flattenClassifierModelCatalog("deepseek", values);
 //# sourceMappingURL=deepseek.models.js.map

package/dist/providers/deepseek.models.js.map

@@ -1 +1 @@
-{"version":3,"file":"deepseek.models.js","sourceRoot":"","sources":["../../src/providers/deepseek.models.ts"],"names":[],"mappings":"AAAA,4DAA4D;AAC5D,iEAAiE;AAEjE,OAAO,MAAM,MAAM,sBAAsB,CAAC,OAAO,IAAI,EAAE,MAAM,EAAE,CAAC;AAChE,OAAO,EAAE,mBAAmB,EAAqB,MAAM,qBAAqB,CAAC;AAE7E,MAAM,CAAC,MAAM,eAAe,GAC3B,mBAAmB,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC","sourcesContent":["// This file is auto-generated by scripts/generate-models.ts\n// Do not edit manually - run 'npm run generate-models' to update\n\nimport values from \"./data/deepseek.json\" with { type: \"json\" };\nimport { flattenModelCatalog, type ModelCatalog } from \"../model-catalog.ts\";\n\nexport const DEEPSEEK_MODELS: ModelCatalog<typeof values, \"deepseek\"> =\n\tflattenModelCatalog(\"deepseek\", values);\n"]}
+{"version":3,"file":"deepseek.models.js","sourceRoot":"","sources":["../../src/providers/deepseek.models.ts"],"names":[],"mappings":"AAAA,4DAA4D;AAC5D,iEAAiE;AAEjE,OAAO,MAAM,MAAM,sBAAsB,CAAC,OAAO,IAAI,EAAE,MAAM,EAAE,CAAC;AAChE,OAAO,EAAE,uBAAuB,EAAE,wBAAwB,EAAE,6BAA6B,EAA8E,MAAM,qBAAqB,CAAC;AAEnM,MAAM,CAAC,MAAM,eAAe,GAC3B,uBAAuB,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;AAE7C,MAAM,CAAC,MAAM,qBAAqB,GACjC,wBAAwB,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;AAE9C,MAAM,CAAC,MAAM,0BAA0B,GACtC,6BAA6B,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC","sourcesContent":["// This file is auto-generated by scripts/generate-models.ts\n// Do not edit manually - run 'npm run generate-models' to update\n\nimport values from \"./data/deepseek.json\" with { type: \"json\" };\nimport { flattenChatModelCatalog, flattenImageModelCatalog, flattenClassifierModelCatalog, type ChatModelCatalog, type ImageModelCatalog, type ClassifierModelCatalog } from \"../model-catalog.ts\";\n\nexport const DEEPSEEK_MODELS: ChatModelCatalog<typeof import(\"./data/deepseek.json\", { with: { type: \"json\" } }), \"deepseek\"> =\n\tflattenChatModelCatalog(\"deepseek\", values);\n\nexport const DEEPSEEK_IMAGE_MODELS: ImageModelCatalog<typeof import(\"./data/deepseek.json\", { with: { type: \"json\" } }), \"deepseek\"> =\n\tflattenImageModelCatalog(\"deepseek\", values);\n\nexport const DEEPSEEK_CLASSIFIER_MODELS: ClassifierModelCatalog<typeof import(\"./data/deepseek.json\", { with: { type: \"json\" } }), \"deepseek\"> =\n\tflattenClassifierModelCatalog(\"deepseek\", values);\n"]}

package/dist/providers/faux.d.ts

@@ -1,10 +1,11 @@
 import { type Provider } from "../models.ts";
-import type { AssistantMessage, AssistantMessageEventStream, Context, DeferredCancelOptions, DeferredFetchOptions, DeferredHandle, Model, SimpleStreamOptions, StreamFunction, TextContent, ThinkingContent, ToolCall } from "../types.ts";
+import type { AssistantMessage, AssistantMessageEventStream, DeferredCancelOptions, DeferredFetchOptions, DeferredHandle, Model, SimpleStreamOptions, StreamFunction, TextContent, ThinkingContent, ToolCall, TranscriptContext } from "../types.ts";
 export interface FauxModelDefinition {
     id: string;
     name?: string;
     reasoning?: boolean;
     input?: ("text" | "image")[];
+    inputLimits?: Model<string>["inputLimits"];
     cost?: {
         input: number;
         output: number;
@@ -32,7 +33,7 @@ export interface FauxProviderState {
     deferredFetchCount: number;
     cancelledDeferred: DeferredHandle[];
 }
-export type FauxResponseFactory = (context: Context, options: SimpleStreamOptions | undefined, state: FauxProviderState, model: Model<string>) => AssistantMessage | Promise<AssistantMessage>;
+export type FauxResponseFactory = (context: TranscriptContext, options: SimpleStreamOptions | undefined, state: FauxProviderState, model: Model<string>) => AssistantMessage | Promise<AssistantMessage>;
 export type FauxResponseStep = AssistantMessage | FauxResponseFactory;
 export interface RegisterFauxProviderOptions {
     api?: string;

package/dist/providers/faux.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"faux.d.ts","sourceRoot":"","sources":["../../src/providers/faux.ts"],"names":[],"mappings":"AAAA,OAAO,EAAkB,KAAK,QAAQ,EAAE,MAAM,cAAc,CAAC;AAC7D,OAAO,KAAK,EACX,gBAAgB,EAChB,2BAA2B,EAC3B,OAAO,EACP,qBAAqB,EACrB,oBAAoB,EACpB,cAAc,EAId,KAAK,EACL,mBAAmB,EACnB,cAAc,EAEd,WAAW,EACX,eAAe,EACf,QAAQ,EAGR,MAAM,aAAa,CAAC;AAoBrB,MAAM,WAAW,mBAAmB;IACnC,EAAE,EAAE,MAAM,CAAC;IACX,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,KAAK,CAAC,EAAE,CAAC,MAAM,GAAG,OAAO,CAAC,EAAE,CAAC;IAC7B,IAAI,CAAC,EAAE;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAC;QAAC,UAAU,EAAE,MAAM,CAAA;KAAE,CAAC;IAChF,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,SAAS,CAAC,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,MAAM,gBAAgB,GAAG,WAAW,GAAG,eAAe,GAAG,QAAQ,CAAC;AAExE,wBAAgB,QAAQ,CAAC,IAAI,EAAE,MAAM,GAAG,WAAW,CAElD;AAED,wBAAgB,YAAY,CAAC,QAAQ,EAAE,MAAM,GAAG,eAAe,CAE9D;AAED,wBAAgB,YAAY,CAAC,IAAI,EAAE,MAAM,EAAE,UAAU,EAAE,QAAQ,CAAC,WAAW,CAAC,EAAE,OAAO,GAAE;IAAE,EAAE,CAAC,EAAE,MAAM,CAAA;CAAO,GAAG,QAAQ,CAOrH;AASD,wBAAgB,oBAAoB,CACnC,OAAO,EAAE,MAAM,GAAG,gBAAgB,GAAG,gBAAgB,EAAE,EACvD,OAAO,GAAE;IACR,UAAU,CAAC,EAAE,gBAAgB,CAAC,YAAY,CAAC,CAAC;IAC5C,QAAQ,CAAC,EAAE,cAAc,CAAC;IAC1B,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,SAAS,CAAC,EAAE,MAAM,CAAC;CACd,GACJ,gBAAgB,CAclB;AAED,MAAM,WAAW,iBAAiB;IACjC,SAAS,EAAE,MAAM,CAAC;IAClB,kBAAkB,EAAE,MAAM,CAAC;IAC3B,iBAAiB,EAAE,cAAc,EAAE,CAAC;CACpC;AAED,MAAM,MAAM,mBAAmB,GAAG,CACjC,OAAO,EAAE,OAAO,EAChB,OAAO,EAAE,mBAAmB,GAAG,SAAS,EACxC,KAAK,EAAE,iBAAiB,EACxB,KAAK,EAAE,KAAK,CAAC,MAAM,CAAC,KAChB,gBAAgB,GAAG,OAAO,CAAC,gBAAgB,CAAC,CAAC;AAElD,MAAM,MAAM,gBAAgB,GAAG,gBAAgB,GAAG,mBAAmB,CAAC;AAEtE,MAAM,WAAW,2BAA2B;IAC3C,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,mBAAmB,EAAE,CAAC;IAC/B,QAAQ,CAAC,EAAE;QACV,oGAAoG;QACpG,cAAc,CAAC,EAAE,MAAM,CAAC;QACxB,WAAW,CAAC,EAAE,MAAM,CAAC;KACrB,CAAC;IACF,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,SAAS,CAAC,EAAE;QACX,GAAG,CAAC,EAAE,MAAM,CAAC;QACb,GAAG,CAAC,EAAE,MAAM,CAAC;KACb,CAAC;CACF;AAED,MAAM,WAAW,wBAAwB;IACxC,GAAG,EAAE,MAAM,CAAC;IACZ,MAAM,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,GAAG,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;IAC5C,QAAQ,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC;IAC1B,QAAQ,CAAC,OAAO,EAAE,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,GAAG,SAAS,CAAC;IACrD,KAAK,EAAE,iBAAiB,CAAC;IACzB,YAAY,EAAE,CAAC,SAAS,EAAE,gBAAgB,EAAE,KAAK,IAAI,CAAC;IACtD,eAAe,EAAE,CAAC,SAAS,EAAE,gBAAgB,EAAE,KAAK,IAAI,CAAC;IACzD,uBAAuB,EAAE,MAAM,MAAM,CAAC;IACtC,UAAU,EAAE,MAAM,IAAI,CAAC;CACvB;AAED,MAAM,WAAW,kBAAkB;IAClC,QAAQ,EAAE,QAAQ,CAAC;IACnB,GAAG,EAAE,MAAM,CAAC;IACZ,MAAM,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,GAAG,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;IAC5C,QAAQ,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC;IAC1B,QAAQ,CAAC,OAAO,EAAE,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,GAAG,SAAS,CAAC;IACrD,KAAK,EAAE,iBAAiB,CAAC;IACzB,YAAY,EAAE,CAAC,SAAS,EAAE,gBAAgB,EAAE,KAAK,IAAI,CAAC;IACtD,eAAe,EAAE,CAAC,SAAS,EAAE,gBAAgB,EAAE,KAAK,IAAI,CAAC;IACzD,uBAAuB,EAAE,MAAM,MAAM,CAAC;CACtC;AA2SD,wBAAgB,cAAc,CAAC,OAAO,EAAE,2BAA2B;;;;;;;;;;;;;;;;EA6OlE;AAED;;;;;;;;;GASG;AACH,wBAAgB,YAAY,CAAC,OAAO,GAAE,2BAAgC,GAAG,kBAAkB,CAuB1F","sourcesContent":["import { createProvider, type Provider } from \"../models.ts\";\nimport type {\n\tAssistantMessage,\n\tAssistantMessageEventStream,\n\tContext,\n\tDeferredCancelOptions,\n\tDeferredFetchOptions,\n\tDeferredHandle,\n\tDocumentContent,\n\tImageContent,\n\tMessage,\n\tModel,\n\tSimpleStreamOptions,\n\tStreamFunction,\n\tStreamOptions,\n\tTextContent,\n\tThinkingContent,\n\tToolCall,\n\tToolResultMessage,\n\tUsage,\n} from \"../types.ts\";\nimport { createAssistantMessageEventStream } from \"../utils/event-stream.ts\";\n\nconst DEFAULT_API = \"faux\";\nconst DEFAULT_PROVIDER = \"faux\";\nconst DEFAULT_MODEL_ID = \"faux-1\";\nconst DEFAULT_MODEL_NAME = \"Faux Model\";\nconst DEFAULT_BASE_URL = \"http://localhost:0\";\nconst DEFAULT_MIN_TOKEN_SIZE = 3;\nconst DEFAULT_MAX_TOKEN_SIZE = 5;\n\nconst DEFAULT_USAGE: Usage = {\n\tinput: 0,\n\toutput: 0,\n\tcacheRead: 0,\n\tcacheWrite: 0,\n\ttotalTokens: 0,\n\tcost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },\n};\n\nexport interface FauxModelDefinition {\n\tid: string;\n\tname?: string;\n\treasoning?: boolean;\n\tinput?: (\"text\" | \"image\")[];\n\tcost?: { input: number; output: number; cacheRead: number; cacheWrite: number };\n\tcontextWindow?: number;\n\tmaxTokens?: number;\n}\n\nexport type FauxContentBlock = TextContent | ThinkingContent | ToolCall;\n\nexport function fauxText(text: string): TextContent {\n\treturn { type: \"text\", text };\n}\n\nexport function fauxThinking(thinking: string): ThinkingContent {\n\treturn { type: \"thinking\", thinking };\n}\n\nexport function fauxToolCall(name: string, arguments_: ToolCall[\"arguments\"], options: { id?: string } = {}): ToolCall {\n\treturn {\n\t\ttype: \"toolCall\",\n\t\tid: options.id ?? randomId(\"tool\"),\n\t\tname,\n\t\targuments: arguments_,\n\t};\n}\n\nfunction normalizeFauxAssistantContent(content: string | FauxContentBlock | FauxContentBlock[]): FauxContentBlock[] {\n\tif (typeof content === \"string\") {\n\t\treturn [fauxText(content)];\n\t}\n\treturn Array.isArray(content) ? content : [content];\n}\n\nexport function fauxAssistantMessage(\n\tcontent: string | FauxContentBlock | FauxContentBlock[],\n\toptions: {\n\t\tstopReason?: AssistantMessage[\"stopReason\"];\n\t\tdeferred?: DeferredHandle;\n\t\terrorMessage?: string;\n\t\tresponseId?: string;\n\t\ttimestamp?: number;\n\t} = {},\n): AssistantMessage {\n\treturn {\n\t\trole: \"assistant\",\n\t\tcontent: normalizeFauxAssistantContent(content),\n\t\tapi: DEFAULT_API,\n\t\tprovider: DEFAULT_PROVIDER,\n\t\tmodel: DEFAULT_MODEL_ID,\n\t\tusage: DEFAULT_USAGE,\n\t\tstopReason: options.stopReason ?? \"stop\",\n\t\tdeferred: options.deferred,\n\t\terrorMessage: options.errorMessage,\n\t\tresponseId: options.responseId,\n\t\ttimestamp: options.timestamp ?? Date.now(),\n\t};\n}\n\nexport interface FauxProviderState {\n\tcallCount: number;\n\tdeferredFetchCount: number;\n\tcancelledDeferred: DeferredHandle[];\n}\n\nexport type FauxResponseFactory = (\n\tcontext: Context,\n\toptions: SimpleStreamOptions | undefined,\n\tstate: FauxProviderState,\n\tmodel: Model<string>,\n) => AssistantMessage | Promise<AssistantMessage>;\n\nexport type FauxResponseStep = AssistantMessage | FauxResponseFactory;\n\nexport interface RegisterFauxProviderOptions {\n\tapi?: string;\n\tprovider?: string;\n\tmodels?: FauxModelDefinition[];\n\tdeferred?: {\n\t\t/** Number of fetches that return the original handle before the scripted response becomes ready. */\n\t\tpendingFetches?: number;\n\t\tpollAfterMs?: number;\n\t};\n\ttokensPerSecond?: number;\n\ttokenSize?: {\n\t\tmin?: number;\n\t\tmax?: number;\n\t};\n}\n\nexport interface FauxProviderRegistration {\n\tapi: string;\n\tmodels: [Model<string>, ...Model<string>[]];\n\tgetModel(): Model<string>;\n\tgetModel(modelId: string): Model<string> | undefined;\n\tstate: FauxProviderState;\n\tsetResponses: (responses: FauxResponseStep[]) => void;\n\tappendResponses: (responses: FauxResponseStep[]) => void;\n\tgetPendingResponseCount: () => number;\n\tunregister: () => void;\n}\n\nexport interface FauxProviderHandle {\n\tprovider: Provider;\n\tapi: string;\n\tmodels: [Model<string>, ...Model<string>[]];\n\tgetModel(): Model<string>;\n\tgetModel(modelId: string): Model<string> | undefined;\n\tstate: FauxProviderState;\n\tsetResponses: (responses: FauxResponseStep[]) => void;\n\tappendResponses: (responses: FauxResponseStep[]) => void;\n\tgetPendingResponseCount: () => number;\n}\n\nfunction estimateTokens(text: string): number {\n\treturn Math.ceil(text.length / 4);\n}\n\nfunction randomId(prefix: string): string {\n\treturn `${prefix}:${Date.now()}:${Math.random().toString(36).slice(2)}`;\n}\n\nfunction contentToText(content: string | Array<TextContent | ImageContent | DocumentContent>): string {\n\tif (typeof content === \"string\") {\n\t\treturn content;\n\t}\n\treturn content\n\t\t.map((block) => {\n\t\t\tif (block.type === \"text\") {\n\t\t\t\treturn block.text;\n\t\t\t}\n\t\t\tif (block.type === \"document\") {\n\t\t\t\treturn `[document:${block.mimeType}:${block.data.length}]`;\n\t\t\t}\n\t\t\treturn `[image:${block.mimeType}:${block.data.length}]`;\n\t\t})\n\t\t.join(\"\\n\");\n}\n\nfunction assistantContentToText(content: AssistantMessage[\"content\"]): string {\n\treturn content\n\t\t.map((block) => {\n\t\t\tif (block.type === \"text\") {\n\t\t\t\treturn block.text;\n\t\t\t}\n\t\t\tif (block.type === \"thinking\") {\n\t\t\t\treturn block.thinking;\n\t\t\t}\n\t\t\t// Fallback boundary markers carry no text.\n\t\t\tif (block.type !== \"toolCall\") {\n\t\t\t\treturn \"\";\n\t\t\t}\n\t\t\treturn `${block.name}:${JSON.stringify(block.arguments)}`;\n\t\t})\n\t\t.join(\"\\n\");\n}\n\nfunction toolResultToText(message: ToolResultMessage): string {\n\treturn [message.toolName, ...message.content.map((block) => contentToText([block]))].join(\"\\n\");\n}\n\nfunction messageToText(message: Message): string {\n\tif (message.role === \"user\") {\n\t\treturn contentToText(message.content);\n\t}\n\tif (message.role === \"assistant\") {\n\t\treturn assistantContentToText(message.content);\n\t}\n\treturn toolResultToText(message);\n}\n\nfunction serializeContext(context: Context): string {\n\tconst parts: string[] = [];\n\tif (context.systemPrompt) {\n\t\tparts.push(`system:${context.systemPrompt}`);\n\t}\n\tfor (const message of context.messages) {\n\t\tparts.push(`${message.role}:${messageToText(message)}`);\n\t}\n\tif (context.tools?.length) {\n\t\tparts.push(`tools:${JSON.stringify(context.tools)}`);\n\t}\n\treturn parts.join(\"\\n\\n\");\n}\n\nfunction commonPrefixLength(a: string, b: string): number {\n\tconst length = Math.min(a.length, b.length);\n\tlet index = 0;\n\twhile (index < length && a[index] === b[index]) {\n\t\tindex++;\n\t}\n\treturn index;\n}\n\nfunction withUsageEstimate(\n\tmessage: AssistantMessage,\n\tcontext: Context,\n\toptions: StreamOptions | undefined,\n\tpromptCache: Map<string, string>,\n): AssistantMessage {\n\tconst promptText = serializeContext(context);\n\tconst promptTokens = estimateTokens(promptText);\n\tconst outputTokens = estimateTokens(assistantContentToText(message.content));\n\tlet input = promptTokens;\n\tlet cacheRead = 0;\n\tlet cacheWrite = 0;\n\tconst sessionId = options?.sessionId;\n\n\tif (sessionId && options?.cacheRetention !== \"none\") {\n\t\tconst previousPrompt = promptCache.get(sessionId);\n\t\tif (previousPrompt) {\n\t\t\tconst cachedChars = commonPrefixLength(previousPrompt, promptText);\n\t\t\tcacheRead = estimateTokens(previousPrompt.slice(0, cachedChars));\n\t\t\tcacheWrite = estimateTokens(promptText.slice(cachedChars));\n\t\t\tinput = Math.max(0, promptTokens - cacheRead);\n\t\t} else {\n\t\t\tcacheWrite = promptTokens;\n\t\t}\n\t\tpromptCache.set(sessionId, promptText);\n\t}\n\n\treturn {\n\t\t...message,\n\t\tusage: {\n\t\t\tinput,\n\t\t\toutput: outputTokens,\n\t\t\tcacheRead,\n\t\t\tcacheWrite,\n\t\t\ttotalTokens: input + outputTokens + cacheRead + cacheWrite,\n\t\t\tcost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },\n\t\t},\n\t};\n}\n\nfunction splitStringByTokenSize(text: string, minTokenSize: number, maxTokenSize: number): string[] {\n\tconst chunks: string[] = [];\n\tlet index = 0;\n\twhile (index < text.length) {\n\t\tconst tokenSize = minTokenSize + Math.floor(Math.random() * (maxTokenSize - minTokenSize + 1));\n\t\tconst charSize = Math.max(1, tokenSize * 4);\n\t\tchunks.push(text.slice(index, index + charSize));\n\t\tindex += charSize;\n\t}\n\treturn chunks.length > 0 ? chunks : [\"\"];\n}\n\nfunction cloneMessage(message: AssistantMessage, api: string, provider: string, modelId: string): AssistantMessage {\n\tconst cloned = structuredClone(message);\n\treturn {\n\t\t...cloned,\n\t\tapi,\n\t\tprovider,\n\t\tmodel: modelId,\n\t\ttimestamp: cloned.timestamp ?? Date.now(),\n\t\tusage: cloned.usage ?? DEFAULT_USAGE,\n\t};\n}\n\nfunction createDeferredMessage(model: Model<string>, handle: DeferredHandle): AssistantMessage {\n\treturn {\n\t\trole: \"assistant\",\n\t\tcontent: [],\n\t\tapi: model.api,\n\t\tprovider: model.provider,\n\t\tmodel: model.id,\n\t\tusage: DEFAULT_USAGE,\n\t\tstopReason: \"deferred\",\n\t\tdeferred: handle,\n\t\ttimestamp: Date.now(),\n\t};\n}\n\nfunction createErrorMessage(error: unknown, api: string, provider: string, modelId: string): AssistantMessage {\n\treturn {\n\t\trole: \"assistant\",\n\t\tcontent: [],\n\t\tapi,\n\t\tprovider,\n\t\tmodel: modelId,\n\t\tusage: DEFAULT_USAGE,\n\t\tstopReason: \"error\",\n\t\terrorMessage: error instanceof Error ? error.message : String(error),\n\t\ttimestamp: Date.now(),\n\t};\n}\n\nfunction createAbortedMessage(partial: AssistantMessage): AssistantMessage {\n\treturn {\n\t\t...partial,\n\t\tstopReason: \"aborted\",\n\t\terrorMessage: \"Request was aborted\",\n\t\ttimestamp: Date.now(),\n\t};\n}\n\nfunction scheduleChunk(chunk: string, tokensPerSecond: number | undefined): Promise<void> {\n\tif (!tokensPerSecond || tokensPerSecond <= 0) {\n\t\treturn new Promise((resolve) => queueMicrotask(resolve));\n\t}\n\tconst delayMs = (estimateTokens(chunk) / tokensPerSecond) * 1000;\n\treturn new Promise((resolve) => setTimeout(resolve, delayMs));\n}\n\nasync function streamWithDeltas(\n\tstream: AssistantMessageEventStream,\n\tmessage: AssistantMessage,\n\tminTokenSize: number,\n\tmaxTokenSize: number,\n\ttokensPerSecond: number | undefined,\n\tsignal: AbortSignal | undefined,\n): Promise<void> {\n\tconst partial: AssistantMessage = { ...message, content: [], stopReason: \"pending\" };\n\tif (signal?.aborted) {\n\t\tconst aborted = createAbortedMessage(partial);\n\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\tstream.end(aborted);\n\t\treturn;\n\t}\n\n\tstream.push({ type: \"start\", partial: { ...partial } });\n\n\tfor (let index = 0; index < message.content.length; index++) {\n\t\tif (signal?.aborted) {\n\t\t\tconst aborted = createAbortedMessage(partial);\n\t\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\t\tstream.end(aborted);\n\t\t\treturn;\n\t\t}\n\n\t\tconst block = message.content[index];\n\n\t\tif (block.type === \"thinking\") {\n\t\t\tpartial.content = [...partial.content, { type: \"thinking\", thinking: \"\" }];\n\t\t\tstream.push({ type: \"thinking_start\", contentIndex: index, partial: { ...partial } });\n\t\t\tfor (const chunk of splitStringByTokenSize(block.thinking, minTokenSize, maxTokenSize)) {\n\t\t\t\tawait scheduleChunk(chunk, tokensPerSecond);\n\t\t\t\tif (signal?.aborted) {\n\t\t\t\t\tconst aborted = createAbortedMessage(partial);\n\t\t\t\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\t\t\t\tstream.end(aborted);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\t\t\t\t(partial.content[index] as ThinkingContent).thinking += chunk;\n\t\t\t\tstream.push({ type: \"thinking_delta\", contentIndex: index, delta: chunk, partial: { ...partial } });\n\t\t\t}\n\t\t\tstream.push({\n\t\t\t\ttype: \"thinking_end\",\n\t\t\t\tcontentIndex: index,\n\t\t\t\tcontent: block.thinking,\n\t\t\t\tpartial: { ...partial },\n\t\t\t});\n\t\t\tcontinue;\n\t\t}\n\n\t\tif (block.type === \"text\") {\n\t\t\tpartial.content = [...partial.content, { type: \"text\", text: \"\" }];\n\t\t\tstream.push({ type: \"text_start\", contentIndex: index, partial: { ...partial } });\n\t\t\tfor (const chunk of splitStringByTokenSize(block.text, minTokenSize, maxTokenSize)) {\n\t\t\t\tawait scheduleChunk(chunk, tokensPerSecond);\n\t\t\t\tif (signal?.aborted) {\n\t\t\t\t\tconst aborted = createAbortedMessage(partial);\n\t\t\t\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\t\t\t\tstream.end(aborted);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\t\t\t\t(partial.content[index] as TextContent).text += chunk;\n\t\t\t\tstream.push({ type: \"text_delta\", contentIndex: index, delta: chunk, partial: { ...partial } });\n\t\t\t}\n\t\t\tstream.push({ type: \"text_end\", contentIndex: index, content: block.text, partial: { ...partial } });\n\t\t\tcontinue;\n\t\t}\n\n\t\t// Fallback boundary markers are replay metadata with nothing to stream, but they still\n\t\t// occupy a slot in `partial.content` so later blocks keep their source `contentIndex`.\n\t\t// This matches the real Anthropic provider, which pushes the wire `fallback` block into\n\t\t// `output.content` and emits no event for it; every `contentIndex` consumer indexes\n\t\t// `partial.content` directly, so skipping the slot would shift every later block by one.\n\t\tif (block.type !== \"toolCall\") {\n\t\t\tpartial.content = [...partial.content, block];\n\t\t\tcontinue;\n\t\t}\n\n\t\tpartial.content = [...partial.content, { type: \"toolCall\", id: block.id, name: block.name, arguments: {} }];\n\t\tstream.push({ type: \"toolcall_start\", contentIndex: index, partial: { ...partial } });\n\t\tfor (const chunk of splitStringByTokenSize(JSON.stringify(block.arguments), minTokenSize, maxTokenSize)) {\n\t\t\tawait scheduleChunk(chunk, tokensPerSecond);\n\t\t\tif (signal?.aborted) {\n\t\t\t\tconst aborted = createAbortedMessage(partial);\n\t\t\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\t\t\tstream.end(aborted);\n\t\t\t\treturn;\n\t\t\t}\n\t\t\tstream.push({ type: \"toolcall_delta\", contentIndex: index, delta: chunk, partial: { ...partial } });\n\t\t}\n\t\t(partial.content[index] as ToolCall).arguments = block.arguments;\n\t\tstream.push({ type: \"toolcall_end\", contentIndex: index, toolCall: block, partial: { ...partial } });\n\t}\n\n\tif (message.stopReason === \"pending\") {\n\t\tthrow new Error(\"Faux response ended without a stop reason\");\n\t}\n\tif (message.stopReason === \"error\" || message.stopReason === \"aborted\") {\n\t\tstream.push({ type: \"error\", reason: message.stopReason, error: message });\n\t\tstream.end(message);\n\t\treturn;\n\t}\n\n\tstream.push({ type: \"done\", reason: message.stopReason, message });\n\tstream.end(message);\n}\n\nexport function createFauxCore(options: RegisterFauxProviderOptions) {\n\tconst api = options.api ?? randomId(DEFAULT_API);\n\tconst provider = options.provider ?? DEFAULT_PROVIDER;\n\tconst minTokenSize = Math.max(\n\t\t1,\n\t\tMath.min(options.tokenSize?.min ?? DEFAULT_MIN_TOKEN_SIZE, options.tokenSize?.max ?? DEFAULT_MAX_TOKEN_SIZE),\n\t);\n\tconst maxTokenSize = Math.max(minTokenSize, options.tokenSize?.max ?? DEFAULT_MAX_TOKEN_SIZE);\n\tlet pendingResponses: FauxResponseStep[] = [];\n\tconst tokensPerSecond = options.tokensPerSecond;\n\tconst state: FauxProviderState = { callCount: 0, deferredFetchCount: 0, cancelledDeferred: [] };\n\tconst promptCache = new Map<string, string>();\n\tconst deferredResponses = new Map<\n\t\tstring,\n\t\t{\n\t\t\thandle: DeferredHandle;\n\t\t\tstep: FauxResponseStep;\n\t\t\tcontext: Context;\n\t\t\toptions: SimpleStreamOptions | undefined;\n\t\t\tmodel: Model<string>;\n\t\t\tpendingFetches: number;\n\t\t\tcancelled: boolean;\n\t\t\tfinal?: AssistantMessage;\n\t\t}\n\t>();\n\n\tconst modelDefinitions = options.models?.length\n\t\t? options.models\n\t\t: [\n\t\t\t\t{\n\t\t\t\t\tid: DEFAULT_MODEL_ID,\n\t\t\t\t\tname: DEFAULT_MODEL_NAME,\n\t\t\t\t\treasoning: false,\n\t\t\t\t\tinput: [\"text\", \"image\"] as (\"text\" | \"image\")[],\n\t\t\t\t\tcost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },\n\t\t\t\t\tcontextWindow: 128000,\n\t\t\t\t\tmaxTokens: 16384,\n\t\t\t\t},\n\t\t\t];\n\tconst models = modelDefinitions.map((definition) => ({\n\t\tid: definition.id,\n\t\tname: definition.name ?? definition.id,\n\t\tapi,\n\t\tprovider,\n\t\tbaseUrl: DEFAULT_BASE_URL,\n\t\treasoning: definition.reasoning ?? false,\n\t\tinput: definition.input ?? [\"text\", \"image\"],\n\t\tcost: definition.cost ?? { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },\n\t\tcontextWindow: definition.contextWindow ?? 128000,\n\t\tmaxTokens: definition.maxTokens ?? 16384,\n\t})) as [Model<string>, ...Model<string>[]];\n\n\tconst resolveResponse = async (\n\t\tstep: FauxResponseStep,\n\t\tcontext: Context,\n\t\tstreamOptions: SimpleStreamOptions | undefined,\n\t\trequestModel: Model<string>,\n\t): Promise<AssistantMessage> => {\n\t\tconst resolved = typeof step === \"function\" ? await step(context, streamOptions, state, requestModel) : step;\n\t\treturn withUsageEstimate(\n\t\t\tcloneMessage(resolved, api, provider, requestModel.id),\n\t\t\tcontext,\n\t\t\tstreamOptions,\n\t\t\tpromptCache,\n\t\t);\n\t};\n\n\tconst stream: StreamFunction<string, SimpleStreamOptions> = (requestModel, context, streamOptions) => {\n\t\tconst outer = createAssistantMessageEventStream();\n\t\tconst step = pendingResponses.shift();\n\t\tstate.callCount++;\n\n\t\tqueueMicrotask(async () => {\n\t\t\ttry {\n\t\t\t\tawait streamOptions?.onResponse?.({ status: 200, headers: {} }, requestModel);\n\t\t\t\tif (!step) {\n\t\t\t\t\tlet message = createErrorMessage(\n\t\t\t\t\t\tnew Error(\"No more faux responses queued\"),\n\t\t\t\t\t\tapi,\n\t\t\t\t\t\tprovider,\n\t\t\t\t\t\trequestModel.id,\n\t\t\t\t\t);\n\t\t\t\t\tmessage = withUsageEstimate(message, context, streamOptions, promptCache);\n\t\t\t\t\touter.push({ type: \"error\", reason: \"error\", error: message });\n\t\t\t\t\touter.end(message);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\n\t\t\t\tif (streamOptions?.deferred) {\n\t\t\t\t\tconst handle: DeferredHandle = {\n\t\t\t\t\t\tprovider: requestModel.provider,\n\t\t\t\t\t\tmodelId: requestModel.id,\n\t\t\t\t\t\tapi: requestModel.api,\n\t\t\t\t\t\tid: randomId(\"deferred\"),\n\t\t\t\t\t\t...(options.deferred?.pollAfterMs !== undefined ? { pollAfterMs: options.deferred.pollAfterMs } : {}),\n\t\t\t\t\t};\n\t\t\t\t\tdeferredResponses.set(handle.id, {\n\t\t\t\t\t\thandle,\n\t\t\t\t\t\tstep,\n\t\t\t\t\t\tcontext,\n\t\t\t\t\t\toptions: streamOptions,\n\t\t\t\t\t\tmodel: requestModel,\n\t\t\t\t\t\tpendingFetches: Math.max(0, Math.floor(options.deferred?.pendingFetches ?? 0)),\n\t\t\t\t\t\tcancelled: false,\n\t\t\t\t\t});\n\t\t\t\t\tawait streamWithDeltas(\n\t\t\t\t\t\touter,\n\t\t\t\t\t\tcreateDeferredMessage(requestModel, handle),\n\t\t\t\t\t\tminTokenSize,\n\t\t\t\t\t\tmaxTokenSize,\n\t\t\t\t\t\ttokensPerSecond,\n\t\t\t\t\t\tstreamOptions.signal,\n\t\t\t\t\t);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\n\t\t\t\tconst message = await resolveResponse(step, context, streamOptions, requestModel);\n\t\t\t\tawait streamWithDeltas(outer, message, minTokenSize, maxTokenSize, tokensPerSecond, streamOptions?.signal);\n\t\t\t} catch (error) {\n\t\t\t\tconst message = createErrorMessage(error, api, provider, requestModel.id);\n\t\t\t\touter.push({ type: \"error\", reason: \"error\", error: message });\n\t\t\t\touter.end(message);\n\t\t\t}\n\t\t});\n\n\t\treturn outer;\n\t};\n\n\tconst streamSimple: StreamFunction<string, SimpleStreamOptions> = (streamModel, context, streamOptions) =>\n\t\tstream(streamModel, context, streamOptions);\n\n\tconst fetchDeferred = (\n\t\trequestModel: Model<string>,\n\t\thandle: DeferredHandle,\n\t\tfetchOptions?: DeferredFetchOptions,\n\t): AssistantMessageEventStream => {\n\t\tconst outer = createAssistantMessageEventStream();\n\t\tstate.deferredFetchCount++;\n\n\t\tqueueMicrotask(async () => {\n\t\t\ttry {\n\t\t\t\tawait fetchOptions?.onResponse?.({ status: 200, headers: {} }, requestModel);\n\t\t\t\tconst entry = deferredResponses.get(handle.id);\n\t\t\t\tif (\n\t\t\t\t\t!entry ||\n\t\t\t\t\tentry.handle.provider !== handle.provider ||\n\t\t\t\t\tentry.handle.modelId !== handle.modelId ||\n\t\t\t\t\tentry.handle.api !== handle.api\n\t\t\t\t) {\n\t\t\t\t\tthrow new Error(`Unknown faux deferred response: ${handle.id}`);\n\t\t\t\t}\n\t\t\t\tif (entry.cancelled) throw new Error(`Faux deferred response was cancelled: ${handle.id}`);\n\n\t\t\t\tif (entry.pendingFetches > 0) {\n\t\t\t\t\tentry.pendingFetches--;\n\t\t\t\t\tawait streamWithDeltas(\n\t\t\t\t\t\touter,\n\t\t\t\t\t\tcreateDeferredMessage(requestModel, entry.handle),\n\t\t\t\t\t\tminTokenSize,\n\t\t\t\t\t\tmaxTokenSize,\n\t\t\t\t\t\ttokensPerSecond,\n\t\t\t\t\t\tfetchOptions?.signal,\n\t\t\t\t\t);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\n\t\t\t\tif (!entry.final) {\n\t\t\t\t\tconst {\n\t\t\t\t\t\tdeferred: _deferred,\n\t\t\t\t\t\tsignal: _submissionSignal,\n\t\t\t\t\t\tonResponse: _submissionOnResponse,\n\t\t\t\t\t\t...submissionOptions\n\t\t\t\t\t} = entry.options ?? {};\n\t\t\t\t\ttry {\n\t\t\t\t\t\tentry.final = await resolveResponse(entry.step, entry.context, submissionOptions, entry.model);\n\t\t\t\t\t} catch (error) {\n\t\t\t\t\t\tentry.final = createErrorMessage(error, api, provider, entry.model.id);\n\t\t\t\t\t}\n\t\t\t\t}\n\t\t\t\tawait streamWithDeltas(\n\t\t\t\t\touter,\n\t\t\t\t\tentry.final,\n\t\t\t\t\tminTokenSize,\n\t\t\t\t\tmaxTokenSize,\n\t\t\t\t\ttokensPerSecond,\n\t\t\t\t\tfetchOptions?.signal,\n\t\t\t\t);\n\t\t\t} catch (error) {\n\t\t\t\tconst message = createErrorMessage(error, api, provider, requestModel.id);\n\t\t\t\touter.push({ type: \"error\", reason: \"error\", error: message });\n\t\t\t\touter.end(message);\n\t\t\t}\n\t\t});\n\n\t\treturn outer;\n\t};\n\n\tconst cancelDeferred = async (\n\t\trequestModel: Model<string>,\n\t\thandle: DeferredHandle,\n\t\tcancelOptions?: DeferredCancelOptions,\n\t): Promise<void> => {\n\t\tstate.cancelledDeferred.push(structuredClone(handle));\n\t\tconst entry = deferredResponses.get(handle.id);\n\t\tif (entry) entry.cancelled = true;\n\t\tawait cancelOptions?.onResponse?.({ status: 200, headers: {} }, requestModel);\n\t};\n\n\tfunction getModel(): Model<string>;\n\tfunction getModel(requestedModelId: string): Model<string> | undefined;\n\tfunction getModel(requestedModelId?: string): Model<string> | undefined {\n\t\tif (!requestedModelId) {\n\t\t\treturn models[0];\n\t\t}\n\t\treturn models.find((candidate) => candidate.id === requestedModelId);\n\t}\n\n\treturn {\n\t\tapi,\n\t\tprovider,\n\t\tmodels,\n\t\tstream,\n\t\tstreamSimple,\n\t\tfetchDeferred,\n\t\tcancelDeferred,\n\t\tgetModel,\n\t\tstate,\n\t\tsetResponses(responses: FauxResponseStep[]) {\n\t\t\tpendingResponses = [...responses];\n\t\t},\n\t\tappendResponses(responses: FauxResponseStep[]) {\n\t\t\tpendingResponses.push(...responses);\n\t\t},\n\t\tgetPendingResponseCount() {\n\t\t\treturn pendingResponses.length;\n\t\t},\n\t};\n}\n\n/**\n * Faux provider for tests built on explicit `Models` collections:\n *\n * ```ts\n * const faux = fauxProvider();\n * const models = createModels();\n * models.setProvider(faux.provider);\n * faux.setResponses([fauxAssistantMessage(\"hi\")]);\n * ```\n */\nexport function fauxProvider(options: RegisterFauxProviderOptions = {}): FauxProviderHandle {\n\tconst core = createFauxCore(options);\n\tconst provider = createProvider({\n\t\tid: core.provider,\n\t\tauth: { apiKey: { name: \"Faux\", resolve: async () => ({ auth: {} }) } },\n\t\tmodels: core.models,\n\t\tapi: {\n\t\t\tstream: core.stream,\n\t\t\tstreamSimple: core.streamSimple,\n\t\t\tfetchDeferred: core.fetchDeferred,\n\t\t\tcancelDeferred: core.cancelDeferred,\n\t\t},\n\t});\n\treturn {\n\t\tprovider,\n\t\tapi: core.api,\n\t\tmodels: core.models,\n\t\tgetModel: core.getModel,\n\t\tstate: core.state,\n\t\tsetResponses: core.setResponses,\n\t\tappendResponses: core.appendResponses,\n\t\tgetPendingResponseCount: core.getPendingResponseCount,\n\t};\n}\n"]}
+{"version":3,"file":"faux.d.ts","sourceRoot":"","sources":["../../src/providers/faux.ts"],"names":[],"mappings":"AAAA,OAAO,EAAkB,KAAK,QAAQ,EAAE,MAAM,cAAc,CAAC;AAC7D,OAAO,KAAK,EACX,gBAAgB,EAChB,2BAA2B,EAC3B,qBAAqB,EACrB,oBAAoB,EACpB,cAAc,EAId,KAAK,EACL,mBAAmB,EACnB,cAAc,EAEd,WAAW,EACX,eAAe,EACf,QAAQ,EAER,iBAAiB,EAEjB,MAAM,aAAa,CAAC;AAqBrB,MAAM,WAAW,mBAAmB;IACnC,EAAE,EAAE,MAAM,CAAC;IACX,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,KAAK,CAAC,EAAE,CAAC,MAAM,GAAG,OAAO,CAAC,EAAE,CAAC;IAC7B,WAAW,CAAC,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC,aAAa,CAAC,CAAC;IAC3C,IAAI,CAAC,EAAE;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAC;QAAC,UAAU,EAAE,MAAM,CAAA;KAAE,CAAC;IAChF,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,SAAS,CAAC,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,MAAM,gBAAgB,GAAG,WAAW,GAAG,eAAe,GAAG,QAAQ,CAAC;AAExE,wBAAgB,QAAQ,CAAC,IAAI,EAAE,MAAM,GAAG,WAAW,CAElD;AAED,wBAAgB,YAAY,CAAC,QAAQ,EAAE,MAAM,GAAG,eAAe,CAE9D;AAED,wBAAgB,YAAY,CAAC,IAAI,EAAE,MAAM,EAAE,UAAU,EAAE,QAAQ,CAAC,WAAW,CAAC,EAAE,OAAO,GAAE;IAAE,EAAE,CAAC,EAAE,MAAM,CAAA;CAAO,GAAG,QAAQ,CAOrH;AASD,wBAAgB,oBAAoB,CACnC,OAAO,EAAE,MAAM,GAAG,gBAAgB,GAAG,gBAAgB,EAAE,EACvD,OAAO,GAAE;IACR,UAAU,CAAC,EAAE,gBAAgB,CAAC,YAAY,CAAC,CAAC;IAC5C,QAAQ,CAAC,EAAE,cAAc,CAAC;IAC1B,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,SAAS,CAAC,EAAE,MAAM,CAAC;CACd,GACJ,gBAAgB,CAclB;AAED,MAAM,WAAW,iBAAiB;IACjC,SAAS,EAAE,MAAM,CAAC;IAClB,kBAAkB,EAAE,MAAM,CAAC;IAC3B,iBAAiB,EAAE,cAAc,EAAE,CAAC;CACpC;AAED,MAAM,MAAM,mBAAmB,GAAG,CACjC,OAAO,EAAE,iBAAiB,EAC1B,OAAO,EAAE,mBAAmB,GAAG,SAAS,EACxC,KAAK,EAAE,iBAAiB,EACxB,KAAK,EAAE,KAAK,CAAC,MAAM,CAAC,KAChB,gBAAgB,GAAG,OAAO,CAAC,gBAAgB,CAAC,CAAC;AAElD,MAAM,MAAM,gBAAgB,GAAG,gBAAgB,GAAG,mBAAmB,CAAC;AAEtE,MAAM,WAAW,2BAA2B;IAC3C,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,mBAAmB,EAAE,CAAC;IAC/B,QAAQ,CAAC,EAAE;QACV,oGAAoG;QACpG,cAAc,CAAC,EAAE,MAAM,CAAC;QACxB,WAAW,CAAC,EAAE,MAAM,CAAC;KACrB,CAAC;IACF,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,SAAS,CAAC,EAAE;QACX,GAAG,CAAC,EAAE,MAAM,CAAC;QACb,GAAG,CAAC,EAAE,MAAM,CAAC;KACb,CAAC;CACF;AAED,MAAM,WAAW,wBAAwB;IACxC,GAAG,EAAE,MAAM,CAAC;IACZ,MAAM,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,GAAG,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;IAC5C,QAAQ,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC;IAC1B,QAAQ,CAAC,OAAO,EAAE,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,GAAG,SAAS,CAAC;IACrD,KAAK,EAAE,iBAAiB,CAAC;IACzB,YAAY,EAAE,CAAC,SAAS,EAAE,gBAAgB,EAAE,KAAK,IAAI,CAAC;IACtD,eAAe,EAAE,CAAC,SAAS,EAAE,gBAAgB,EAAE,KAAK,IAAI,CAAC;IACzD,uBAAuB,EAAE,MAAM,MAAM,CAAC;IACtC,UAAU,EAAE,MAAM,IAAI,CAAC;CACvB;AAED,MAAM,WAAW,kBAAkB;IAClC,QAAQ,EAAE,QAAQ,CAAC;IACnB,GAAG,EAAE,MAAM,CAAC;IACZ,MAAM,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,GAAG,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;IAC5C,QAAQ,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC;IAC1B,QAAQ,CAAC,OAAO,EAAE,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,GAAG,SAAS,CAAC;IACrD,KAAK,EAAE,iBAAiB,CAAC;IACzB,YAAY,EAAE,CAAC,SAAS,EAAE,gBAAgB,EAAE,KAAK,IAAI,CAAC;IACtD,eAAe,EAAE,CAAC,SAAS,EAAE,gBAAgB,EAAE,KAAK,IAAI,CAAC;IACzD,uBAAuB,EAAE,MAAM,MAAM,CAAC;CACtC;AA0SD,wBAAgB,cAAc,CAAC,OAAO,EAAE,2BAA2B;;;;;;;;;;;;;;;;EA8OlE;AAED;;;;;;;;;GASG;AACH,wBAAgB,YAAY,CAAC,OAAO,GAAE,2BAAgC,GAAG,kBAAkB,CAuB1F","sourcesContent":["import { createProvider, type Provider } from \"../models.ts\";\nimport type {\n\tAssistantMessage,\n\tAssistantMessageEventStream,\n\tDeferredCancelOptions,\n\tDeferredFetchOptions,\n\tDeferredHandle,\n\tDocumentContent,\n\tImageContent,\n\tMessage,\n\tModel,\n\tSimpleStreamOptions,\n\tStreamFunction,\n\tStreamOptions,\n\tTextContent,\n\tThinkingContent,\n\tToolCall,\n\tToolResultMessage,\n\tTranscriptContext,\n\tUsage,\n} from \"../types.ts\";\nimport { createAssistantMessageEventStream } from \"../utils/event-stream.ts\";\nimport { getSystemMessageText } from \"../utils/text.ts\";\n\nconst DEFAULT_API = \"faux\";\nconst DEFAULT_PROVIDER = \"faux\";\nconst DEFAULT_MODEL_ID = \"faux-1\";\nconst DEFAULT_MODEL_NAME = \"Faux Model\";\nconst DEFAULT_BASE_URL = \"http://localhost:0\";\nconst DEFAULT_MIN_TOKEN_SIZE = 3;\nconst DEFAULT_MAX_TOKEN_SIZE = 5;\n\nconst DEFAULT_USAGE: Usage = {\n\tinput: 0,\n\toutput: 0,\n\tcacheRead: 0,\n\tcacheWrite: 0,\n\ttotalTokens: 0,\n\tcost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },\n};\n\nexport interface FauxModelDefinition {\n\tid: string;\n\tname?: string;\n\treasoning?: boolean;\n\tinput?: (\"text\" | \"image\")[];\n\tinputLimits?: Model<string>[\"inputLimits\"];\n\tcost?: { input: number; output: number; cacheRead: number; cacheWrite: number };\n\tcontextWindow?: number;\n\tmaxTokens?: number;\n}\n\nexport type FauxContentBlock = TextContent | ThinkingContent | ToolCall;\n\nexport function fauxText(text: string): TextContent {\n\treturn { type: \"text\", text };\n}\n\nexport function fauxThinking(thinking: string): ThinkingContent {\n\treturn { type: \"thinking\", thinking };\n}\n\nexport function fauxToolCall(name: string, arguments_: ToolCall[\"arguments\"], options: { id?: string } = {}): ToolCall {\n\treturn {\n\t\ttype: \"toolCall\",\n\t\tid: options.id ?? randomId(\"tool\"),\n\t\tname,\n\t\targuments: arguments_,\n\t};\n}\n\nfunction normalizeFauxAssistantContent(content: string | FauxContentBlock | FauxContentBlock[]): FauxContentBlock[] {\n\tif (typeof content === \"string\") {\n\t\treturn [fauxText(content)];\n\t}\n\treturn Array.isArray(content) ? content : [content];\n}\n\nexport function fauxAssistantMessage(\n\tcontent: string | FauxContentBlock | FauxContentBlock[],\n\toptions: {\n\t\tstopReason?: AssistantMessage[\"stopReason\"];\n\t\tdeferred?: DeferredHandle;\n\t\terrorMessage?: string;\n\t\tresponseId?: string;\n\t\ttimestamp?: number;\n\t} = {},\n): AssistantMessage {\n\treturn {\n\t\trole: \"assistant\",\n\t\tcontent: normalizeFauxAssistantContent(content),\n\t\tapi: DEFAULT_API,\n\t\tprovider: DEFAULT_PROVIDER,\n\t\tmodel: DEFAULT_MODEL_ID,\n\t\tusage: DEFAULT_USAGE,\n\t\tstopReason: options.stopReason ?? \"stop\",\n\t\tdeferred: options.deferred,\n\t\terrorMessage: options.errorMessage,\n\t\tresponseId: options.responseId,\n\t\ttimestamp: options.timestamp ?? Date.now(),\n\t};\n}\n\nexport interface FauxProviderState {\n\tcallCount: number;\n\tdeferredFetchCount: number;\n\tcancelledDeferred: DeferredHandle[];\n}\n\nexport type FauxResponseFactory = (\n\tcontext: TranscriptContext,\n\toptions: SimpleStreamOptions | undefined,\n\tstate: FauxProviderState,\n\tmodel: Model<string>,\n) => AssistantMessage | Promise<AssistantMessage>;\n\nexport type FauxResponseStep = AssistantMessage | FauxResponseFactory;\n\nexport interface RegisterFauxProviderOptions {\n\tapi?: string;\n\tprovider?: string;\n\tmodels?: FauxModelDefinition[];\n\tdeferred?: {\n\t\t/** Number of fetches that return the original handle before the scripted response becomes ready. */\n\t\tpendingFetches?: number;\n\t\tpollAfterMs?: number;\n\t};\n\ttokensPerSecond?: number;\n\ttokenSize?: {\n\t\tmin?: number;\n\t\tmax?: number;\n\t};\n}\n\nexport interface FauxProviderRegistration {\n\tapi: string;\n\tmodels: [Model<string>, ...Model<string>[]];\n\tgetModel(): Model<string>;\n\tgetModel(modelId: string): Model<string> | undefined;\n\tstate: FauxProviderState;\n\tsetResponses: (responses: FauxResponseStep[]) => void;\n\tappendResponses: (responses: FauxResponseStep[]) => void;\n\tgetPendingResponseCount: () => number;\n\tunregister: () => void;\n}\n\nexport interface FauxProviderHandle {\n\tprovider: Provider;\n\tapi: string;\n\tmodels: [Model<string>, ...Model<string>[]];\n\tgetModel(): Model<string>;\n\tgetModel(modelId: string): Model<string> | undefined;\n\tstate: FauxProviderState;\n\tsetResponses: (responses: FauxResponseStep[]) => void;\n\tappendResponses: (responses: FauxResponseStep[]) => void;\n\tgetPendingResponseCount: () => number;\n}\n\nfunction estimateTokens(text: string): number {\n\treturn Math.ceil(text.length / 4);\n}\n\nfunction randomId(prefix: string): string {\n\treturn `${prefix}:${Date.now()}:${Math.random().toString(36).slice(2)}`;\n}\n\nfunction contentToText(content: string | Array<TextContent | ImageContent | DocumentContent>): string {\n\tif (typeof content === \"string\") {\n\t\treturn content;\n\t}\n\treturn content\n\t\t.map((block) => {\n\t\t\tif (block.type === \"text\") {\n\t\t\t\treturn block.text;\n\t\t\t}\n\t\t\tif (block.type === \"document\") {\n\t\t\t\treturn `[document:${block.mimeType}:${block.data.length}]`;\n\t\t\t}\n\t\t\treturn `[image:${block.mimeType}:${block.data.length}]`;\n\t\t})\n\t\t.join(\"\\n\");\n}\n\nfunction assistantContentToText(content: AssistantMessage[\"content\"]): string {\n\treturn content\n\t\t.map((block) => {\n\t\t\tif (block.type === \"text\") {\n\t\t\t\treturn block.text;\n\t\t\t}\n\t\t\tif (block.type === \"thinking\") {\n\t\t\t\treturn block.thinking;\n\t\t\t}\n\t\t\t// Fallback boundary markers carry no text.\n\t\t\tif (block.type !== \"toolCall\") {\n\t\t\t\treturn \"\";\n\t\t\t}\n\t\t\treturn `${block.name}:${JSON.stringify(block.arguments)}`;\n\t\t})\n\t\t.join(\"\\n\");\n}\n\nfunction toolResultToText(message: ToolResultMessage): string {\n\treturn [message.toolName, ...message.content.map((block) => contentToText([block]))].join(\"\\n\");\n}\n\nfunction messageToText(message: Message): string {\n\tif (message.role === \"system\") {\n\t\treturn [\n\t\t\tgetSystemMessageText(message),\n\t\t\t...(message.toolsRemoved?.map((tool) => `tool-:${JSON.stringify(tool)}`) ?? []),\n\t\t\t...(message.toolsAdded?.map((tool) => `tool+:${JSON.stringify(tool)}`) ?? []),\n\t\t]\n\t\t\t.filter((part) => part.length > 0)\n\t\t\t.join(\"\\n\");\n\t}\n\tif (message.role === \"user\") {\n\t\treturn contentToText(message.content);\n\t}\n\tif (message.role === \"assistant\") {\n\t\treturn assistantContentToText(message.content);\n\t}\n\treturn toolResultToText(message);\n}\n\nfunction serializeContext(context: TranscriptContext): string {\n\treturn context.messages.map((message) => `${message.role}:${messageToText(message)}`).join(\"\\n\\n\");\n}\n\nfunction commonPrefixLength(a: string, b: string): number {\n\tconst length = Math.min(a.length, b.length);\n\tlet index = 0;\n\twhile (index < length && a[index] === b[index]) {\n\t\tindex++;\n\t}\n\treturn index;\n}\n\nfunction withUsageEstimate(\n\tmessage: AssistantMessage,\n\tcontext: TranscriptContext,\n\toptions: StreamOptions | undefined,\n\tpromptCache: Map<string, string>,\n): AssistantMessage {\n\tconst promptText = serializeContext(context);\n\tconst promptTokens = estimateTokens(promptText);\n\tconst outputTokens = estimateTokens(assistantContentToText(message.content));\n\tlet input = promptTokens;\n\tlet cacheRead = 0;\n\tlet cacheWrite = 0;\n\tconst sessionId = options?.sessionId;\n\n\tif (sessionId && options?.cacheRetention !== \"none\") {\n\t\tconst previousPrompt = promptCache.get(sessionId);\n\t\tif (previousPrompt) {\n\t\t\tconst cachedChars = commonPrefixLength(previousPrompt, promptText);\n\t\t\tcacheRead = estimateTokens(previousPrompt.slice(0, cachedChars));\n\t\t\tcacheWrite = estimateTokens(promptText.slice(cachedChars));\n\t\t\tinput = Math.max(0, promptTokens - cacheRead);\n\t\t} else {\n\t\t\tcacheWrite = promptTokens;\n\t\t}\n\t\tpromptCache.set(sessionId, promptText);\n\t}\n\n\treturn {\n\t\t...message,\n\t\tusage: {\n\t\t\tinput,\n\t\t\toutput: outputTokens,\n\t\t\tcacheRead,\n\t\t\tcacheWrite,\n\t\t\ttotalTokens: input + outputTokens + cacheRead + cacheWrite,\n\t\t\tcost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },\n\t\t},\n\t};\n}\n\nfunction splitStringByTokenSize(text: string, minTokenSize: number, maxTokenSize: number): string[] {\n\tconst chunks: string[] = [];\n\tlet index = 0;\n\twhile (index < text.length) {\n\t\tconst tokenSize = minTokenSize + Math.floor(Math.random() * (maxTokenSize - minTokenSize + 1));\n\t\tconst charSize = Math.max(1, tokenSize * 4);\n\t\tchunks.push(text.slice(index, index + charSize));\n\t\tindex += charSize;\n\t}\n\treturn chunks.length > 0 ? chunks : [\"\"];\n}\n\nfunction cloneMessage(message: AssistantMessage, api: string, provider: string, modelId: string): AssistantMessage {\n\tconst cloned = structuredClone(message);\n\treturn {\n\t\t...cloned,\n\t\tapi,\n\t\tprovider,\n\t\tmodel: modelId,\n\t\ttimestamp: cloned.timestamp ?? Date.now(),\n\t\tusage: cloned.usage ?? DEFAULT_USAGE,\n\t};\n}\n\nfunction createDeferredMessage(model: Model<string>, handle: DeferredHandle): AssistantMessage {\n\treturn {\n\t\trole: \"assistant\",\n\t\tcontent: [],\n\t\tapi: model.api,\n\t\tprovider: model.provider,\n\t\tmodel: model.id,\n\t\tusage: DEFAULT_USAGE,\n\t\tstopReason: \"deferred\",\n\t\tdeferred: handle,\n\t\ttimestamp: Date.now(),\n\t};\n}\n\nfunction createErrorMessage(error: unknown, api: string, provider: string, modelId: string): AssistantMessage {\n\treturn {\n\t\trole: \"assistant\",\n\t\tcontent: [],\n\t\tapi,\n\t\tprovider,\n\t\tmodel: modelId,\n\t\tusage: DEFAULT_USAGE,\n\t\tstopReason: \"error\",\n\t\terrorMessage: error instanceof Error ? error.message : String(error),\n\t\ttimestamp: Date.now(),\n\t};\n}\n\nfunction createAbortedMessage(partial: AssistantMessage): AssistantMessage {\n\treturn {\n\t\t...partial,\n\t\tstopReason: \"aborted\",\n\t\terrorMessage: \"Request was aborted\",\n\t\ttimestamp: Date.now(),\n\t};\n}\n\nfunction scheduleChunk(chunk: string, tokensPerSecond: number | undefined): Promise<void> {\n\tif (!tokensPerSecond || tokensPerSecond <= 0) {\n\t\treturn new Promise((resolve) => queueMicrotask(resolve));\n\t}\n\tconst delayMs = (estimateTokens(chunk) / tokensPerSecond) * 1000;\n\treturn new Promise((resolve) => setTimeout(resolve, delayMs));\n}\n\nasync function streamWithDeltas(\n\tstream: AssistantMessageEventStream,\n\tmessage: AssistantMessage,\n\tminTokenSize: number,\n\tmaxTokenSize: number,\n\ttokensPerSecond: number | undefined,\n\tsignal: AbortSignal | undefined,\n): Promise<void> {\n\tconst partial: AssistantMessage = { ...message, content: [], stopReason: \"pending\" };\n\tif (signal?.aborted) {\n\t\tconst aborted = createAbortedMessage(partial);\n\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\tstream.end(aborted);\n\t\treturn;\n\t}\n\n\tstream.push({ type: \"start\", partial: { ...partial } });\n\n\tfor (let index = 0; index < message.content.length; index++) {\n\t\tif (signal?.aborted) {\n\t\t\tconst aborted = createAbortedMessage(partial);\n\t\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\t\tstream.end(aborted);\n\t\t\treturn;\n\t\t}\n\n\t\tconst block = message.content[index];\n\n\t\tif (block.type === \"thinking\") {\n\t\t\tpartial.content = [...partial.content, { type: \"thinking\", thinking: \"\" }];\n\t\t\tstream.push({ type: \"thinking_start\", contentIndex: index, partial: { ...partial } });\n\t\t\tfor (const chunk of splitStringByTokenSize(block.thinking, minTokenSize, maxTokenSize)) {\n\t\t\t\tawait scheduleChunk(chunk, tokensPerSecond);\n\t\t\t\tif (signal?.aborted) {\n\t\t\t\t\tconst aborted = createAbortedMessage(partial);\n\t\t\t\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\t\t\t\tstream.end(aborted);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\t\t\t\t(partial.content[index] as ThinkingContent).thinking += chunk;\n\t\t\t\tstream.push({ type: \"thinking_delta\", contentIndex: index, delta: chunk, partial: { ...partial } });\n\t\t\t}\n\t\t\tstream.push({\n\t\t\t\ttype: \"thinking_end\",\n\t\t\t\tcontentIndex: index,\n\t\t\t\tcontent: block.thinking,\n\t\t\t\tpartial: { ...partial },\n\t\t\t});\n\t\t\tcontinue;\n\t\t}\n\n\t\tif (block.type === \"text\") {\n\t\t\tpartial.content = [...partial.content, { type: \"text\", text: \"\" }];\n\t\t\tstream.push({ type: \"text_start\", contentIndex: index, partial: { ...partial } });\n\t\t\tfor (const chunk of splitStringByTokenSize(block.text, minTokenSize, maxTokenSize)) {\n\t\t\t\tawait scheduleChunk(chunk, tokensPerSecond);\n\t\t\t\tif (signal?.aborted) {\n\t\t\t\t\tconst aborted = createAbortedMessage(partial);\n\t\t\t\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\t\t\t\tstream.end(aborted);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\t\t\t\t(partial.content[index] as TextContent).text += chunk;\n\t\t\t\tstream.push({ type: \"text_delta\", contentIndex: index, delta: chunk, partial: { ...partial } });\n\t\t\t}\n\t\t\tstream.push({ type: \"text_end\", contentIndex: index, content: block.text, partial: { ...partial } });\n\t\t\tcontinue;\n\t\t}\n\n\t\t// Fallback boundary markers are replay metadata with nothing to stream, but they still\n\t\t// occupy a slot in `partial.content` so later blocks keep their source `contentIndex`.\n\t\t// This matches the real Anthropic provider, which pushes the wire `fallback` block into\n\t\t// `output.content` and emits no event for it; every `contentIndex` consumer indexes\n\t\t// `partial.content` directly, so skipping the slot would shift every later block by one.\n\t\tif (block.type !== \"toolCall\") {\n\t\t\tpartial.content = [...partial.content, block];\n\t\t\tcontinue;\n\t\t}\n\n\t\tpartial.content = [...partial.content, { type: \"toolCall\", id: block.id, name: block.name, arguments: {} }];\n\t\tstream.push({ type: \"toolcall_start\", contentIndex: index, partial: { ...partial } });\n\t\tfor (const chunk of splitStringByTokenSize(JSON.stringify(block.arguments), minTokenSize, maxTokenSize)) {\n\t\t\tawait scheduleChunk(chunk, tokensPerSecond);\n\t\t\tif (signal?.aborted) {\n\t\t\t\tconst aborted = createAbortedMessage(partial);\n\t\t\t\tstream.push({ type: \"error\", reason: \"aborted\", error: aborted });\n\t\t\t\tstream.end(aborted);\n\t\t\t\treturn;\n\t\t\t}\n\t\t\tstream.push({ type: \"toolcall_delta\", contentIndex: index, delta: chunk, partial: { ...partial } });\n\t\t}\n\t\t(partial.content[index] as ToolCall).arguments = block.arguments;\n\t\tstream.push({ type: \"toolcall_end\", contentIndex: index, toolCall: block, partial: { ...partial } });\n\t}\n\n\tif (message.stopReason === \"pending\") {\n\t\tthrow new Error(\"Faux response ended without a stop reason\");\n\t}\n\tif (message.stopReason === \"error\" || message.stopReason === \"aborted\") {\n\t\tstream.push({ type: \"error\", reason: message.stopReason, error: message });\n\t\tstream.end(message);\n\t\treturn;\n\t}\n\n\tstream.push({ type: \"done\", reason: message.stopReason, message });\n\tstream.end(message);\n}\n\nexport function createFauxCore(options: RegisterFauxProviderOptions) {\n\tconst api = options.api ?? randomId(DEFAULT_API);\n\tconst provider = options.provider ?? DEFAULT_PROVIDER;\n\tconst minTokenSize = Math.max(\n\t\t1,\n\t\tMath.min(options.tokenSize?.min ?? DEFAULT_MIN_TOKEN_SIZE, options.tokenSize?.max ?? DEFAULT_MAX_TOKEN_SIZE),\n\t);\n\tconst maxTokenSize = Math.max(minTokenSize, options.tokenSize?.max ?? DEFAULT_MAX_TOKEN_SIZE);\n\tlet pendingResponses: FauxResponseStep[] = [];\n\tconst tokensPerSecond = options.tokensPerSecond;\n\tconst state: FauxProviderState = { callCount: 0, deferredFetchCount: 0, cancelledDeferred: [] };\n\tconst promptCache = new Map<string, string>();\n\tconst deferredResponses = new Map<\n\t\tstring,\n\t\t{\n\t\t\thandle: DeferredHandle;\n\t\t\tstep: FauxResponseStep;\n\t\t\tcontext: TranscriptContext;\n\t\t\toptions: SimpleStreamOptions | undefined;\n\t\t\tmodel: Model<string>;\n\t\t\tpendingFetches: number;\n\t\t\tcancelled: boolean;\n\t\t\tfinal?: AssistantMessage;\n\t\t}\n\t>();\n\n\tconst modelDefinitions = options.models?.length\n\t\t? options.models\n\t\t: [\n\t\t\t\t{\n\t\t\t\t\tid: DEFAULT_MODEL_ID,\n\t\t\t\t\tname: DEFAULT_MODEL_NAME,\n\t\t\t\t\treasoning: false,\n\t\t\t\t\tinput: [\"text\", \"image\"] as (\"text\" | \"image\")[],\n\t\t\t\t\tcost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },\n\t\t\t\t\tcontextWindow: 128000,\n\t\t\t\t\tmaxTokens: 16384,\n\t\t\t\t},\n\t\t\t];\n\tconst models = modelDefinitions.map((definition) => ({\n\t\tid: definition.id,\n\t\tname: definition.name ?? definition.id,\n\t\tapi,\n\t\tprovider,\n\t\tbaseUrl: DEFAULT_BASE_URL,\n\t\treasoning: definition.reasoning ?? false,\n\t\tinput: definition.input ?? [\"text\", \"image\"],\n\t\tinputLimits: definition.inputLimits,\n\t\tcost: definition.cost ?? { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },\n\t\tcontextWindow: definition.contextWindow ?? 128000,\n\t\tmaxTokens: definition.maxTokens ?? 16384,\n\t})) as [Model<string>, ...Model<string>[]];\n\n\tconst resolveResponse = async (\n\t\tstep: FauxResponseStep,\n\t\tcontext: TranscriptContext,\n\t\tstreamOptions: SimpleStreamOptions | undefined,\n\t\trequestModel: Model<string>,\n\t): Promise<AssistantMessage> => {\n\t\tconst resolved = typeof step === \"function\" ? await step(context, streamOptions, state, requestModel) : step;\n\t\treturn withUsageEstimate(\n\t\t\tcloneMessage(resolved, api, provider, requestModel.id),\n\t\t\tcontext,\n\t\t\tstreamOptions,\n\t\t\tpromptCache,\n\t\t);\n\t};\n\n\tconst stream: StreamFunction<string, SimpleStreamOptions> = (requestModel, context, streamOptions) => {\n\t\tconst outer = createAssistantMessageEventStream();\n\t\tconst step = pendingResponses.shift();\n\t\tstate.callCount++;\n\n\t\tqueueMicrotask(async () => {\n\t\t\ttry {\n\t\t\t\tawait streamOptions?.onResponse?.({ status: 200, headers: {} }, requestModel);\n\t\t\t\tif (!step) {\n\t\t\t\t\tlet message = createErrorMessage(\n\t\t\t\t\t\tnew Error(\"No more faux responses queued\"),\n\t\t\t\t\t\tapi,\n\t\t\t\t\t\tprovider,\n\t\t\t\t\t\trequestModel.id,\n\t\t\t\t\t);\n\t\t\t\t\tmessage = withUsageEstimate(message, context, streamOptions, promptCache);\n\t\t\t\t\touter.push({ type: \"error\", reason: \"error\", error: message });\n\t\t\t\t\touter.end(message);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\n\t\t\t\tif (streamOptions?.deferred) {\n\t\t\t\t\tconst handle: DeferredHandle = {\n\t\t\t\t\t\tprovider: requestModel.provider,\n\t\t\t\t\t\tmodelId: requestModel.id,\n\t\t\t\t\t\tapi: requestModel.api,\n\t\t\t\t\t\tid: randomId(\"deferred\"),\n\t\t\t\t\t\t...(options.deferred?.pollAfterMs !== undefined ? { pollAfterMs: options.deferred.pollAfterMs } : {}),\n\t\t\t\t\t};\n\t\t\t\t\tdeferredResponses.set(handle.id, {\n\t\t\t\t\t\thandle,\n\t\t\t\t\t\tstep,\n\t\t\t\t\t\tcontext,\n\t\t\t\t\t\toptions: streamOptions,\n\t\t\t\t\t\tmodel: requestModel,\n\t\t\t\t\t\tpendingFetches: Math.max(0, Math.floor(options.deferred?.pendingFetches ?? 0)),\n\t\t\t\t\t\tcancelled: false,\n\t\t\t\t\t});\n\t\t\t\t\tawait streamWithDeltas(\n\t\t\t\t\t\touter,\n\t\t\t\t\t\tcreateDeferredMessage(requestModel, handle),\n\t\t\t\t\t\tminTokenSize,\n\t\t\t\t\t\tmaxTokenSize,\n\t\t\t\t\t\ttokensPerSecond,\n\t\t\t\t\t\tstreamOptions.signal,\n\t\t\t\t\t);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\n\t\t\t\tconst message = await resolveResponse(step, context, streamOptions, requestModel);\n\t\t\t\tawait streamWithDeltas(outer, message, minTokenSize, maxTokenSize, tokensPerSecond, streamOptions?.signal);\n\t\t\t} catch (error) {\n\t\t\t\tconst message = createErrorMessage(error, api, provider, requestModel.id);\n\t\t\t\touter.push({ type: \"error\", reason: \"error\", error: message });\n\t\t\t\touter.end(message);\n\t\t\t}\n\t\t});\n\n\t\treturn outer;\n\t};\n\n\tconst streamSimple: StreamFunction<string, SimpleStreamOptions> = (streamModel, context, streamOptions) =>\n\t\tstream(streamModel, context, streamOptions);\n\n\tconst fetchDeferred = (\n\t\trequestModel: Model<string>,\n\t\thandle: DeferredHandle,\n\t\tfetchOptions?: DeferredFetchOptions,\n\t): AssistantMessageEventStream => {\n\t\tconst outer = createAssistantMessageEventStream();\n\t\tstate.deferredFetchCount++;\n\n\t\tqueueMicrotask(async () => {\n\t\t\ttry {\n\t\t\t\tawait fetchOptions?.onResponse?.({ status: 200, headers: {} }, requestModel);\n\t\t\t\tconst entry = deferredResponses.get(handle.id);\n\t\t\t\tif (\n\t\t\t\t\t!entry ||\n\t\t\t\t\tentry.handle.provider !== handle.provider ||\n\t\t\t\t\tentry.handle.modelId !== handle.modelId ||\n\t\t\t\t\tentry.handle.api !== handle.api\n\t\t\t\t) {\n\t\t\t\t\tthrow new Error(`Unknown faux deferred response: ${handle.id}`);\n\t\t\t\t}\n\t\t\t\tif (entry.cancelled) throw new Error(`Faux deferred response was cancelled: ${handle.id}`);\n\n\t\t\t\tif (entry.pendingFetches > 0) {\n\t\t\t\t\tentry.pendingFetches--;\n\t\t\t\t\tawait streamWithDeltas(\n\t\t\t\t\t\touter,\n\t\t\t\t\t\tcreateDeferredMessage(requestModel, entry.handle),\n\t\t\t\t\t\tminTokenSize,\n\t\t\t\t\t\tmaxTokenSize,\n\t\t\t\t\t\ttokensPerSecond,\n\t\t\t\t\t\tfetchOptions?.signal,\n\t\t\t\t\t);\n\t\t\t\t\treturn;\n\t\t\t\t}\n\n\t\t\t\tif (!entry.final) {\n\t\t\t\t\tconst {\n\t\t\t\t\t\tdeferred: _deferred,\n\t\t\t\t\t\tsignal: _submissionSignal,\n\t\t\t\t\t\tonResponse: _submissionOnResponse,\n\t\t\t\t\t\t...submissionOptions\n\t\t\t\t\t} = entry.options ?? {};\n\t\t\t\t\ttry {\n\t\t\t\t\t\tentry.final = await resolveResponse(entry.step, entry.context, submissionOptions, entry.model);\n\t\t\t\t\t} catch (error) {\n\t\t\t\t\t\tentry.final = createErrorMessage(error, api, provider, entry.model.id);\n\t\t\t\t\t}\n\t\t\t\t}\n\t\t\t\tawait streamWithDeltas(\n\t\t\t\t\touter,\n\t\t\t\t\tentry.final,\n\t\t\t\t\tminTokenSize,\n\t\t\t\t\tmaxTokenSize,\n\t\t\t\t\ttokensPerSecond,\n\t\t\t\t\tfetchOptions?.signal,\n\t\t\t\t);\n\t\t\t} catch (error) {\n\t\t\t\tconst message = createErrorMessage(error, api, provider, requestModel.id);\n\t\t\t\touter.push({ type: \"error\", reason: \"error\", error: message });\n\t\t\t\touter.end(message);\n\t\t\t}\n\t\t});\n\n\t\treturn outer;\n\t};\n\n\tconst cancelDeferred = async (\n\t\trequestModel: Model<string>,\n\t\thandle: DeferredHandle,\n\t\tcancelOptions?: DeferredCancelOptions,\n\t): Promise<void> => {\n\t\tstate.cancelledDeferred.push(structuredClone(handle));\n\t\tconst entry = deferredResponses.get(handle.id);\n\t\tif (entry) entry.cancelled = true;\n\t\tawait cancelOptions?.onResponse?.({ status: 200, headers: {} }, requestModel);\n\t};\n\n\tfunction getModel(): Model<string>;\n\tfunction getModel(requestedModelId: string): Model<string> | undefined;\n\tfunction getModel(requestedModelId?: string): Model<string> | undefined {\n\t\tif (!requestedModelId) {\n\t\t\treturn models[0];\n\t\t}\n\t\treturn models.find((candidate) => candidate.id === requestedModelId);\n\t}\n\n\treturn {\n\t\tapi,\n\t\tprovider,\n\t\tmodels,\n\t\tstream,\n\t\tstreamSimple,\n\t\tfetchDeferred,\n\t\tcancelDeferred,\n\t\tgetModel,\n\t\tstate,\n\t\tsetResponses(responses: FauxResponseStep[]) {\n\t\t\tpendingResponses = [...responses];\n\t\t},\n\t\tappendResponses(responses: FauxResponseStep[]) {\n\t\t\tpendingResponses.push(...responses);\n\t\t},\n\t\tgetPendingResponseCount() {\n\t\t\treturn pendingResponses.length;\n\t\t},\n\t};\n}\n\n/**\n * Faux provider for tests built on explicit `Models` collections:\n *\n * ```ts\n * const faux = fauxProvider();\n * const models = createModels();\n * models.setProvider(faux.provider);\n * faux.setResponses([fauxAssistantMessage(\"hi\")]);\n * ```\n */\nexport function fauxProvider(options: RegisterFauxProviderOptions = {}): FauxProviderHandle {\n\tconst core = createFauxCore(options);\n\tconst provider = createProvider({\n\t\tid: core.provider,\n\t\tauth: { apiKey: { name: \"Faux\", resolve: async () => ({ auth: {} }) } },\n\t\tmodels: core.models,\n\t\tapi: {\n\t\t\tstream: core.stream,\n\t\t\tstreamSimple: core.streamSimple,\n\t\t\tfetchDeferred: core.fetchDeferred,\n\t\t\tcancelDeferred: core.cancelDeferred,\n\t\t},\n\t});\n\treturn {\n\t\tprovider,\n\t\tapi: core.api,\n\t\tmodels: core.models,\n\t\tgetModel: core.getModel,\n\t\tstate: core.state,\n\t\tsetResponses: core.setResponses,\n\t\tappendResponses: core.appendResponses,\n\t\tgetPendingResponseCount: core.getPendingResponseCount,\n\t};\n}\n"]}

package/dist/providers/faux.js

@@ -1,5 +1,6 @@
 import { createProvider } from "../models.js";
 import { createAssistantMessageEventStream } from "../utils/event-stream.js";
+import { getSystemMessageText } from "../utils/text.js";
 const DEFAULT_API = "faux";
 const DEFAULT_PROVIDER = "faux";
 const DEFAULT_MODEL_ID = "faux-1";
@@ -93,6 +94,15 @@ function toolResultToText(message) {
     return [message.toolName, ...message.content.map((block) => contentToText([block]))].join("\n");
 }
 function messageToText(message) {
+    if (message.role === "system") {
+        return [
+            getSystemMessageText(message),
+            ...(message.toolsRemoved?.map((tool) => `tool-:${JSON.stringify(tool)}`) ?? []),
+            ...(message.toolsAdded?.map((tool) => `tool+:${JSON.stringify(tool)}`) ?? []),
+        ]
+            .filter((part) => part.length > 0)
+            .join("\n");
+    }
     if (message.role === "user") {
         return contentToText(message.content);
     }
@@ -102,17 +112,7 @@ function messageToText(message) {
     return toolResultToText(message);
 }
 function serializeContext(context) {
-    const parts = [];
-    if (context.systemPrompt) {
-        parts.push(`system:${context.systemPrompt}`);
-    }
-    for (const message of context.messages) {
-        parts.push(`${message.role}:${messageToText(message)}`);
-    }
-    if (context.tools?.length) {
-        parts.push(`tools:${JSON.stringify(context.tools)}`);
-    }
-    return parts.join("\n\n");
+    return context.messages.map((message) => `${message.role}:${messageToText(message)}`).join("\n\n");
 }
 function commonPrefixLength(a, b) {
     const length = Math.min(a.length, b.length);
@@ -340,6 +340,7 @@ export function createFauxCore(options) {
         baseUrl: DEFAULT_BASE_URL,
         reasoning: definition.reasoning ?? false,
         input: definition.input ?? ["text", "image"],
+        inputLimits: definition.inputLimits,
         cost: definition.cost ?? { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
         contextWindow: definition.contextWindow ?? 128000,
         maxTokens: definition.maxTokens ?? 16384,