nexus-ai-pro 0.2.0

package/dist/types/providers.js

@@ -0,0 +1,479 @@
+function model(capabilities) {
+    return capabilities;
+}
+const openaiReasoning = {
+    efforts: ['none', 'low', 'medium', 'high', 'xhigh'],
+};
+const gpt5 = (family, maxContextTokens, inputPerMillion, outputPerMillion, qualityScore, speedScore, release, knowledgeCutoff) => model({
+    provider: 'openai',
+    family,
+    modalities: ['text', 'vision'],
+    streaming: true,
+    toolCalling: true,
+    structuredOutputs: true,
+    jsonMode: true,
+    reasoning: openaiReasoning,
+    maxContextTokens,
+    maxOutputTokens: 128000,
+    costPer1kInput: inputPerMillion / 1000,
+    costPer1kOutput: outputPerMillion / 1000,
+    qualityScore,
+    speedScore,
+    release,
+    knowledgeCutoff,
+    status: 'stable',
+    endpoints: ['chat', 'responses'],
+});
+const claude = (family, inputPerMillion, outputPerMillion, qualityScore, speedScore, maxOutputTokens, release, knowledgeCutoff, maxContextTokens = 200000) => model({
+    provider: 'anthropic',
+    family,
+    modalities: ['text', 'vision'],
+    streaming: true,
+    toolCalling: true,
+    structuredOutputs: false,
+    jsonMode: false,
+    reasoning: family.includes('4') || family.includes('3.7'),
+    maxContextTokens,
+    maxOutputTokens,
+    costPer1kInput: inputPerMillion / 1000,
+    costPer1kOutput: outputPerMillion / 1000,
+    qualityScore,
+    speedScore,
+    release,
+    knowledgeCutoff,
+    status: 'stable',
+    endpoints: ['messages'],
+});
+const gemini = (family, maxContextTokens, inputPerMillion, outputPerMillion, qualityScore, speedScore, release, knowledgeCutoff, status = 'stable') => model({
+    provider: 'google',
+    family,
+    modalities: ['text', 'vision', 'audio', 'video', 'pdf'],
+    streaming: true,
+    toolCalling: true,
+    structuredOutputs: true,
+    jsonMode: true,
+    reasoning: true,
+    maxContextTokens,
+    maxOutputTokens: 65536,
+    costPer1kInput: inputPerMillion / 1000,
+    costPer1kOutput: outputPerMillion / 1000,
+    qualityScore,
+    speedScore,
+    release,
+    knowledgeCutoff,
+    status,
+    endpoints: ['generateContent'],
+});
+const openAiCompatible = (provider, family, modalities, maxContextTokens, maxOutputTokens, inputPerMillion, outputPerMillion, qualityScore, speedScore, release, status = 'stable', notes) => model({
+    provider,
+    family,
+    modalities,
+    streaming: true,
+    toolCalling: true,
+    structuredOutputs: true,
+    jsonMode: true,
+    reasoning: family.includes('reasoning') || family.includes('gpt-oss') || family.includes('magistral') || family.includes('gemini-3'),
+    maxContextTokens,
+    maxOutputTokens,
+    costPer1kInput: inputPerMillion / 1000,
+    costPer1kOutput: outputPerMillion / 1000,
+    qualityScore,
+    speedScore,
+    release,
+    status,
+    endpoints: ['chat'],
+    notes,
+});
+const cohere = (family, modalities, maxContextTokens, maxOutputTokens, qualityScore, speedScore, release, notes) => model({
+    provider: 'cohere',
+    family,
+    modalities,
+    streaming: false,
+    toolCalling: true,
+    structuredOutputs: false,
+    jsonMode: false,
+    reasoning: family.includes('reasoning'),
+    maxContextTokens,
+    maxOutputTokens,
+    costPer1kInput: 0,
+    costPer1kOutput: 0,
+    qualityScore,
+    speedScore,
+    release,
+    status: 'stable',
+    endpoints: ['chat'],
+    notes: notes || 'Cohere pricing is deployment/plan dependent; override costs in models.registry for exact estimates.',
+});
+export const KNOWN_MODELS = {
+    // OpenAI - current GPT-5 family
+    'gpt-5.5': gpt5('gpt-5.5', 1000000, 5, 30, 99, 82, '2026', '2025-12-01'),
+    'gpt-5.5-pro': {
+        ...gpt5('gpt-5.5', 1050000, 30, 180, 100, 55, '2026', '2025-12-01'),
+        structuredOutputs: false,
+        endpoints: ['responses'],
+    },
+    'gpt-5.4': gpt5('gpt-5.4', 1050000, 2.5, 15, 97, 85, '2026-03', '2025-08-31'),
+    'gpt-5.4-pro': {
+        ...gpt5('gpt-5.4', 1050000, 30, 180, 99, 55, '2026-03', '2025-08-31'),
+        structuredOutputs: false,
+        endpoints: ['responses'],
+    },
+    'gpt-5.4-mini': gpt5('gpt-5.4', 400000, 0.75, 4.5, 92, 94, '2026-03', '2025-08-31'),
+    'gpt-5.4-nano': gpt5('gpt-5.4', 400000, 0.2, 1.25, 84, 98, '2026-03', '2025-08-31'),
+    'gpt-5.2': gpt5('gpt-5.2', 400000, 1.75, 14, 96, 82, '2025', '2025-08-31'),
+    'gpt-5.2-pro': {
+        ...gpt5('gpt-5.2', 400000, 30, 180, 98, 55, '2025', '2025-08-31'),
+        endpoints: ['responses'],
+    },
+    'gpt-5.2-codex': {
+        ...gpt5('gpt-5.2', 400000, 1.75, 14, 96, 78, '2025', '2025-08-31'),
+        family: 'codex',
+        endpoints: ['responses'],
+    },
+    'gpt-5.1': gpt5('gpt-5.1', 400000, 1.25, 10, 94, 82, '2025', '2024-09-30'),
+    'gpt-5': gpt5('gpt-5', 400000, 1.25, 10, 92, 80, '2025', '2024-09-30'),
+    'gpt-5-mini': gpt5('gpt-5', 400000, 0.25, 2, 88, 94, '2025', '2024-05-31'),
+    'gpt-5-nano': gpt5('gpt-5', 400000, 0.05, 0.4, 78, 99, '2025', '2024-05-31'),
+    'gpt-5-codex': {
+        ...gpt5('gpt-5', 400000, 1.25, 10, 93, 78, '2025', '2024-09-30'),
+        family: 'codex',
+        endpoints: ['responses'],
+    },
+    'gpt-5-chat-latest': {
+        ...gpt5('gpt-5', 128000, 1.25, 10, 90, 82, 'latest', '2024-09-30'),
+        maxOutputTokens: 16384,
+        status: 'latest',
+    },
+    // OpenAI - GPT-4 / reasoning legacy still commonly deployed
+    'gpt-4.1': model({
+        provider: 'openai',
+        family: 'gpt-4.1',
+        modalities: ['text', 'vision'],
+        streaming: true,
+        toolCalling: true,
+        structuredOutputs: true,
+        jsonMode: true,
+        maxContextTokens: 1047576,
+        maxOutputTokens: 32768,
+        costPer1kInput: 0.002,
+        costPer1kOutput: 0.008,
+        qualityScore: 88,
+        speedScore: 80,
+        release: '2025',
+        status: 'stable',
+        endpoints: ['chat', 'responses'],
+    }),
+    'gpt-4.1-mini': model({
+        provider: 'openai',
+        family: 'gpt-4.1',
+        modalities: ['text', 'vision'],
+        streaming: true,
+        toolCalling: true,
+        structuredOutputs: true,
+        jsonMode: true,
+        maxContextTokens: 1047576,
+        maxOutputTokens: 32768,
+        costPer1kInput: 0.0004,
+        costPer1kOutput: 0.0016,
+        qualityScore: 82,
+        speedScore: 94,
+        release: '2025',
+        status: 'stable',
+        endpoints: ['chat', 'responses'],
+    }),
+    'gpt-4.1-nano': model({
+        provider: 'openai',
+        family: 'gpt-4.1',
+        modalities: ['text', 'vision'],
+        streaming: true,
+        toolCalling: true,
+        structuredOutputs: true,
+        jsonMode: true,
+        maxContextTokens: 1047576,
+        maxOutputTokens: 32768,
+        costPer1kInput: 0.0001,
+        costPer1kOutput: 0.0004,
+        qualityScore: 74,
+        speedScore: 98,
+        release: '2025',
+        status: 'stable',
+        endpoints: ['chat', 'responses'],
+    }),
+    'gpt-4o': model({
+        provider: 'openai',
+        family: 'gpt-4o',
+        modalities: ['text', 'vision', 'audio'],
+        streaming: true,
+        toolCalling: true,
+        structuredOutputs: true,
+        jsonMode: true,
+        maxContextTokens: 128000,
+        maxOutputTokens: 16384,
+        costPer1kInput: 0.0025,
+        costPer1kOutput: 0.01,
+        qualityScore: 84,
+        speedScore: 84,
+        release: '2024',
+        status: 'stable',
+        endpoints: ['chat', 'responses', 'realtime'],
+    }),
+    'gpt-4o-mini': model({
+        provider: 'openai',
+        family: 'gpt-4o',
+        modalities: ['text', 'vision'],
+        streaming: true,
+        toolCalling: true,
+        structuredOutputs: true,
+        jsonMode: true,
+        maxContextTokens: 128000,
+        maxOutputTokens: 16384,
+        costPer1kInput: 0.00015,
+        costPer1kOutput: 0.0006,
+        qualityScore: 76,
+        speedScore: 96,
+        release: '2024',
+        status: 'stable',
+        endpoints: ['chat', 'responses'],
+    }),
+    'o3': model({
+        provider: 'openai',
+        family: 'o-series',
+        modalities: ['text', 'vision'],
+        streaming: true,
+        toolCalling: true,
+        structuredOutputs: true,
+        jsonMode: true,
+        reasoning: { efforts: ['low', 'medium', 'high'] },
+        maxContextTokens: 200000,
+        maxOutputTokens: 100000,
+        costPer1kInput: 0.002,
+        costPer1kOutput: 0.008,
+        qualityScore: 90,
+        speedScore: 66,
+        release: '2025',
+        status: 'stable',
+        endpoints: ['chat', 'responses'],
+    }),
+    'o3-mini': model({
+        provider: 'openai',
+        family: 'o-series',
+        modalities: ['text'],
+        streaming: true,
+        toolCalling: true,
+        structuredOutputs: true,
+        jsonMode: true,
+        reasoning: { efforts: ['low', 'medium', 'high'] },
+        maxContextTokens: 200000,
+        maxOutputTokens: 100000,
+        costPer1kInput: 0.0011,
+        costPer1kOutput: 0.0044,
+        qualityScore: 84,
+        speedScore: 82,
+        release: '2025',
+        status: 'stable',
+        endpoints: ['chat', 'responses'],
+    }),
+    'o4-mini': model({
+        provider: 'openai',
+        family: 'o-series',
+        modalities: ['text', 'vision'],
+        streaming: true,
+        toolCalling: true,
+        structuredOutputs: true,
+        jsonMode: true,
+        reasoning: { efforts: ['low', 'medium', 'high'] },
+        maxContextTokens: 200000,
+        maxOutputTokens: 100000,
+        costPer1kInput: 0.0011,
+        costPer1kOutput: 0.0044,
+        qualityScore: 86,
+        speedScore: 88,
+        release: '2025',
+        status: 'stable',
+        endpoints: ['chat', 'responses'],
+    }),
+    'gpt-oss-120b': model({
+        provider: 'openai',
+        family: 'gpt-oss',
+        modalities: ['text'],
+        streaming: true,
+        toolCalling: true,
+        maxContextTokens: 131000,
+        maxOutputTokens: 32768,
+        costPer1kInput: 0,
+        costPer1kOutput: 0,
+        qualityScore: 78,
+        speedScore: 70,
+        release: '2025',
+        status: 'stable',
+        endpoints: [],
+    }),
+    'gpt-oss-20b': model({
+        provider: 'openai',
+        family: 'gpt-oss',
+        modalities: ['text'],
+        streaming: true,
+        toolCalling: true,
+        maxContextTokens: 131000,
+        maxOutputTokens: 32768,
+        costPer1kInput: 0,
+        costPer1kOutput: 0,
+        qualityScore: 68,
+        speedScore: 88,
+        release: '2025',
+        status: 'stable',
+        endpoints: [],
+    }),
+    // Anthropic
+    'claude-opus-4-7': claude('claude-opus-4.7', 5, 25, 99, 72, 128000, '2026', '2026-01', 1000000),
+    'claude-opus-4-6': {
+        ...claude('claude-opus-4.6', 5, 25, 98, 72, 128000, '2026', '2026-01', 1000000),
+        notes: 'Anthropic docs reference claude-opus-4-6 for Claude Code and batch beta usage; verify account and platform availability.',
+    },
+    'claude-opus-4-5-20251101': {
+        ...claude('claude-opus-4.5', 5, 25, 97, 72, 128000, '2025-11-01', '2025-08', 1000000),
+        notes: 'Anthropic Claude Code Bedrock docs reference this Opus 4.5 snapshot; verify direct Claude API availability before production use.',
+    },
+    'claude-sonnet-4-6': claude('claude-sonnet-4.6', 3, 15, 96, 88, 64000, '2026', '2025-08', 1000000),
+    'claude-haiku-4-5-20251001': claude('claude-haiku-4.5', 1, 5, 86, 98, 64000, '2025-10-01', '2025-02'),
+    'claude-opus-4-1-20250805': claude('claude-opus-4.1', 15, 75, 97, 70, 32000, '2025-08-05', '2025-03'),
+    'claude-opus-4-20250514': claude('claude-opus-4', 15, 75, 95, 70, 32000, '2025-05-14', '2025-03'),
+    'claude-sonnet-4-20250514': claude('claude-sonnet-4', 3, 15, 92, 86, 64000, '2025-05-14', '2025-03'),
+    'claude-3-7-sonnet-20250219': claude('claude-3.7', 3, 15, 89, 84, 64000, '2025-02-19', '2024-10'),
+    'claude-3-5-sonnet-20241022': claude('claude-3.5', 3, 15, 86, 84, 8192, '2024-10-22', '2024-04'),
+    'claude-3-5-sonnet-20240620': claude('claude-3.5', 3, 15, 84, 82, 8192, '2024-06-20', '2024-04'),
+    'claude-3-5-haiku-20241022': claude('claude-haiku-3.5', 0.8, 4, 78, 96, 8192, '2024-10-22', '2024-07'),
+    'claude-3-haiku-20240307': claude('claude-haiku-3', 0.25, 1.25, 68, 94, 4096, '2024-03-07', '2023-08'),
+    // Google Gemini
+    'gemini-3.1-pro-preview': gemini('gemini-3.1', 1048576, 2, 12, 97, 76, '2026-02', '2025-01', 'preview'),
+    'gemini-3.1-pro-preview-customtools': gemini('gemini-3.1', 1048576, 2, 12, 97, 74, '2026-02', '2025-01', 'preview'),
+    'gemini-3.1-flash-lite-preview': gemini('gemini-3.1', 1048576, 0.1, 0.4, 86, 98, '2026', '2025-01', 'preview'),
+    'gemini-3-pro-preview': {
+        ...gemini('gemini-3', 1048576, 2, 12, 96, 76, '2025-11', '2025-01', 'preview'),
+        status: 'deprecated',
+        notes: 'Deprecated and shut down by Google on 2026-03-09; migrate to gemini-3.1-pro-preview.',
+    },
+    'gemini-3-flash-preview': gemini('gemini-3', 1048576, 0.5, 3, 90, 94, '2025-12', '2025-01', 'preview'),
+    'gemini-3-pro-image-preview': {
+        ...gemini('gemini-3-image', 65536, 2, 12, 88, 70, '2025-11', '2025-01', 'preview'),
+        modalities: ['text', 'vision', 'image'],
+        toolCalling: false,
+    },
+    'gemini-2.5-pro': gemini('gemini-2.5', 1048576, 1.25, 10, 92, 76, '2025-06', '2025-01'),
+    'gemini-2.5-flash': gemini('gemini-2.5', 1048576, 0.3, 2.5, 86, 94, '2025', '2025-01'),
+    'gemini-2.5-flash-lite': gemini('gemini-2.5', 1048576, 0.1, 0.4, 76, 98, '2025', '2025-01'),
+    'gemini-2.5-flash-lite-preview-09-2025': gemini('gemini-2.5', 1048576, 0.1, 0.4, 76, 98, '2025-09', '2025-01', 'preview'),
+    'gemini-2.0-flash': gemini('gemini-2.0', 1048576, 0.1, 0.4, 78, 94, '2025-02', '2024-08'),
+    'gemini-1.5-pro': {
+        ...gemini('gemini-1.5', 2000000, 1.25, 5, 80, 70, '2024', '2024'),
+        status: 'deprecated',
+    },
+    'gemini-1.5-flash': {
+        ...gemini('gemini-1.5', 1000000, 0.075, 0.3, 72, 90, '2024', '2024'),
+        status: 'deprecated',
+    },
+    // Groq hosted models and systems
+    'groq/openai/gpt-oss-120b': openAiCompatible('groq', 'gpt-oss', ['text'], 131072, 65536, 0.15, 0.6, 86, 96, '2025'),
+    'groq/openai/gpt-oss-20b': openAiCompatible('groq', 'gpt-oss', ['text'], 131072, 65536, 0.075, 0.3, 78, 100, '2025'),
+    'groq/llama-3.3-70b-versatile': openAiCompatible('groq', 'llama-3.3', ['text'], 131072, 32768, 0.59, 0.79, 80, 92, '2024'),
+    'groq/llama-3.1-8b-instant': openAiCompatible('groq', 'llama-3.1', ['text'], 131072, 131072, 0.05, 0.08, 68, 98, '2024'),
+    'groq/groq/compound': openAiCompatible('groq', 'compound', ['text'], 131072, 8192, 0, 0, 82, 94, '2026', 'stable', 'Groq Compound is a hosted agentic system with built-in tools; pricing is not token-metered in the model table.'),
+    'groq/groq/compound-mini': openAiCompatible('groq', 'compound', ['text'], 131072, 8192, 0, 0, 76, 96, '2026', 'stable', 'Groq Compound Mini is a hosted agentic system with built-in tools; pricing is not token-metered in the model table.'),
+    'groq/meta-llama/llama-4-scout-17b-16e-instruct': openAiCompatible('groq', 'llama-4', ['text', 'vision'], 131072, 8192, 0.11, 0.34, 78, 97, '2026', 'preview'),
+    'groq/qwen/qwen3-32b': openAiCompatible('groq', 'qwen3', ['text'], 131072, 40960, 0.29, 0.59, 76, 90, '2026', 'preview'),
+    'groq/openai/gpt-oss-safeguard-20b': openAiCompatible('groq', 'safety', ['text'], 131072, 65536, 0.075, 0.3, 70, 100, '2025', 'preview'),
+    'groq/meta-llama/llama-prompt-guard-2-22m': openAiCompatible('groq', 'prompt-guard', ['text'], 512, 512, 0.03, 0.03, 55, 100, '2026', 'preview'),
+    'groq/meta-llama/llama-prompt-guard-2-86m': openAiCompatible('groq', 'prompt-guard', ['text'], 512, 512, 0.04, 0.04, 58, 100, '2026', 'preview'),
+    // Mistral AI
+    'mistral/mistral-medium-3-5': openAiCompatible('mistral', 'mistral-medium-3.5', ['text', 'vision', 'pdf'], 256000, 8192, 1.5, 7.5, 90, 78, '2026-04'),
+    'mistral/mistral-small-2603': openAiCompatible('mistral', 'mistral-small-4', ['text', 'vision'], 256000, 8192, 0.2, 0.6, 82, 92, '2026-03', 'stable', 'Pricing varies by deployment; override costs in models.registry if needed.'),
+    'mistral/mistral-large-2512': openAiCompatible('mistral', 'mistral-large-3', ['text', 'vision'], 256000, 8192, 2, 6, 88, 76, '2025-12', 'stable', 'Pricing varies by deployment; override costs in models.registry if needed.'),
+    'mistral/devstral-2512': openAiCompatible('mistral', 'devstral-2', ['text'], 256000, 8192, 0.4, 2, 86, 84, '2025-12', 'stable', 'Code-agent model.'),
+    'mistral/codestral-2508': openAiCompatible('mistral', 'codestral', ['text'], 256000, 8192, 0.3, 0.9, 84, 86, '2025-08', 'stable', 'Code model. Pricing varies by deployment; override costs in models.registry if needed.'),
+    'mistral/magistral-medium-2509': openAiCompatible('mistral', 'magistral-medium-1.2', ['text', 'vision'], 128000, 8192, 2, 5, 88, 70, '2025-09', 'stable', 'Reasoning model. Pricing varies by deployment; override costs in models.registry if needed.'),
+    'mistral/mistral-moderation-2603': openAiCompatible('mistral', 'moderation', ['text'], 128000, 4096, 0, 0, 70, 92, '2026-03', 'stable', 'Moderation model, not intended for normal chat completions.'),
+    // Cohere
+    'cohere/command-a-03-2025': cohere('command-a', ['text'], 256000, 8000, 86, 80, '2025-03'),
+    'cohere/command-a-reasoning-08-2025': cohere('command-a-reasoning', ['text'], 256000, 32000, 88, 68, '2025-08'),
+    'cohere/command-a-vision-07-2025': cohere('command-a-vision', ['text', 'vision'], 128000, 8000, 84, 78, '2025-07'),
+    'cohere/command-a-translate-08-2025': cohere('command-a-translate', ['text'], 8000, 8000, 84, 82, '2025-08'),
+    'cohere/command-r7b-12-2024': cohere('command-r7b', ['text'], 128000, 4000, 74, 94, '2024-12'),
+    'cohere/c4ai-aya-expanse-32b': cohere('aya-expanse', ['text'], 128000, 4000, 76, 84, '2024'),
+    'cohere/c4ai-aya-vision-32b': cohere('aya-vision', ['text', 'vision'], 16000, 4000, 78, 78, '2024'),
+    'cohere/tiny-aya-global': cohere('tiny-aya', ['text'], 8000, 8000, 62, 98, '2026'),
+};
+export const MODEL_ALIASES = {
+    'openai/best': 'gpt-5.5',
+    'openai/pro': 'gpt-5.5-pro',
+    'openai/balanced': 'gpt-5.4',
+    'openai/fast': 'gpt-5.4-mini',
+    'openai/cheap': 'gpt-5.4-nano',
+    'openai/coding': 'gpt-5.5',
+    'openai/codex': 'gpt-5-codex',
+    'anthropic/best': 'claude-opus-4-7',
+    'anthropic/balanced': 'claude-sonnet-4-6',
+    'anthropic/fast': 'claude-haiku-4-5-20251001',
+    'google/best': 'gemini-3.1-pro-preview',
+    'google/balanced': 'gemini-2.5-pro',
+    'google/fast': 'gemini-3-flash-preview',
+    'google/cheap': 'gemini-2.5-flash-lite',
+    'groq/best': 'groq/openai/gpt-oss-120b',
+    'groq/fast': 'groq/openai/gpt-oss-20b',
+    'groq/cheap': 'groq/llama-3.1-8b-instant',
+    'groq/compound': 'groq/groq/compound',
+    'mistral/best': 'mistral/mistral-medium-3-5',
+    'mistral/balanced': 'mistral/mistral-large-2512',
+    'mistral/fast': 'mistral/mistral-small-2603',
+    'mistral/coding': 'mistral/devstral-2512',
+    'mistral/mistral-small-4': 'mistral/mistral-small-2603',
+    'mistral/mistral-large-3': 'mistral/mistral-large-2512',
+    'mistral/devstral-2': 'mistral/devstral-2512',
+    'cohere/best': 'cohere/command-a-03-2025',
+    'cohere/reasoning': 'cohere/command-a-reasoning-08-2025',
+    'cohere/vision': 'cohere/command-a-vision-07-2025',
+    'cohere/fast': 'cohere/command-r7b-12-2024',
+    'claude-opus-4.7': 'claude-opus-4-7',
+    'claude-opus-4-7-latest': 'claude-opus-4-7',
+    'claude-opus-4.6': 'claude-opus-4-6',
+    'claude-opus-4-6-latest': 'claude-opus-4-6',
+    'claude-opus-4.5': 'claude-opus-4-5-20251101',
+    'claude-opus-4-5': 'claude-opus-4-5-20251101',
+    'claude-opus-4-5-latest': 'claude-opus-4-5-20251101',
+    'claude-sonnet-4.6': 'claude-sonnet-4-6',
+    'claude-sonnet-4-6-latest': 'claude-sonnet-4-6',
+    'claude-haiku-4.5': 'claude-haiku-4-5-20251001',
+    'claude-haiku-4-5': 'claude-haiku-4-5-20251001',
+    'claude-haiku-4-5-latest': 'claude-haiku-4-5-20251001',
+    'claude-opus-4.1': 'claude-opus-4-1-20250805',
+    'claude-opus-4-1': 'claude-opus-4-1-20250805',
+    'claude-opus-4': 'claude-opus-4-20250514',
+    'claude-opus-4-0': 'claude-opus-4-20250514',
+    'claude-sonnet-4': 'claude-sonnet-4-20250514',
+    'claude-sonnet-4-0': 'claude-sonnet-4-20250514',
+    'claude-sonnet-3.7': 'claude-3-7-sonnet-20250219',
+    'claude-3-7-sonnet-latest': 'claude-3-7-sonnet-20250219',
+    'claude-sonnet-3.5': 'claude-3-5-sonnet-20241022',
+    'claude-3-5-sonnet-latest': 'claude-3-5-sonnet-20241022',
+    'claude-haiku-3.5': 'claude-3-5-haiku-20241022',
+    'claude-3-5-haiku-latest': 'claude-3-5-haiku-20241022',
+    'gemini-pro-latest': 'gemini-3.1-pro-preview',
+    'gemini-flash-latest': 'gemini-3-flash-preview',
+    'gemini-flash-lite-latest': 'gemini-3.1-flash-lite-preview',
+};
+export function resolveProvider(model) {
+    if (model.startsWith('gpt-')
+        || model.startsWith('o1')
+        || model.startsWith('o3')
+        || model.startsWith('o4'))
+        return 'openai';
+    if (model.startsWith('claude-'))
+        return 'anthropic';
+    if (model.startsWith('gemini-'))
+        return 'google';
+    if (model.includes('/')) {
+        const prefix = model.split('/')[0];
+        if (['ollama', 'google', 'groq', 'lmstudio', 'openrouter', 'together', 'deepseek', 'mistral', 'cohere'].includes(prefix))
+            return prefix;
+    }
+    return null;
+}

package/dist/types/response.d.ts

@@ -0,0 +1,55 @@
+import type { PipelineTrace } from '../pipeline/types.js';
+export interface ResponseMeta {
+    requestId: string;
+    providerUsed: string;
+    modelUsed: string;
+    latencyMs: number;
+    tokensInput: number;
+    tokensOutput: number;
+    tokensSaved: number;
+    estimatedCost: string;
+    cacheHit: boolean;
+    guardrailsApplied: string[];
+    routingDecision?: {
+        reason: string;
+        fallbacksConsidered: number;
+    };
+    verification?: {
+        ok: boolean;
+        supportRatio: number;
+        factsChecked: number;
+        unsupportedFacts: string[];
+    };
+    pipeline?: PipelineTrace;
+    semanticCache?: {
+        hit: boolean;
+        score: number;
+        key: string;
+    };
+}
+export interface ToolCall {
+    id: string;
+    type: 'function';
+    function: {
+        name: string;
+        arguments: string;
+    };
+}
+export interface NexusResponse {
+    content: string;
+    role: 'assistant';
+    toolCalls?: ToolCall[];
+    finishReason: 'stop' | 'tool_calls' | 'length' | 'content_filter' | 'error';
+    meta: ResponseMeta;
+}
+export interface StreamChunk {
+    type: 'text' | 'tool_call' | 'done' | 'error';
+    content?: string;
+    toolCall?: ToolCall;
+    meta?: Partial<ResponseMeta>;
+    error?: string;
+}
+export interface NexusStream extends AsyncIterable<StreamChunk> {
+    [Symbol.asyncIterator](): AsyncIterator<StreamChunk>;
+    abort(): void;
+}

package/dist/types/response.js

@@ -0,0 +1 @@
+export {};

package/dist/types/security.d.ts

@@ -0,0 +1,87 @@
+export type SecurityLevel = 'off' | 'basic' | 'standard' | 'strict' | 'paranoid';
+export type SecurityPreset = 'developer' | 'startup' | 'enterprise' | 'healthcare' | 'finance';
+export type SecurityAction = 'allow' | 'block' | 'mask' | 'flag';
+export type PIIType = 'email' | 'phone' | 'credit-card' | 'ip-address' | 'aws-key' | 'private-key';
+export interface InjectionDetectionConfig {
+    enabled?: boolean;
+    sensitivity?: number;
+    onDetection?: 'block' | 'flag' | 'transform';
+    customPatterns?: RegExp[];
+    semantic?: {
+        enabled?: boolean;
+        threshold?: number;
+        examples?: string[];
+    };
+}
+export interface PIIConfig {
+    enabled?: boolean;
+    detect?: PIIType[];
+    action?: 'mask' | 'remove' | 'block' | 'flag';
+    maskChar?: string;
+    preserveFormat?: boolean;
+}
+export interface SecretDetectionConfig {
+    enabled?: boolean;
+    action?: 'block' | 'mask' | 'flag';
+}
+export interface UrlRiskConfig {
+    enabled?: boolean;
+    action?: 'block' | 'flag';
+}
+export interface ToolPolicyConfig {
+    allowedNames?: string[];
+    requiresApproval?: string[];
+}
+export interface ModerationConfig {
+    enabled?: boolean;
+    forbiddenTerms?: string[];
+    onViolation?: 'block' | 'flag' | 'redact';
+}
+export interface DLPConfig {
+    enabled?: boolean;
+    proprietaryTerms?: string[];
+    connectionStrings?: boolean;
+}
+export interface TopicGuardrailConfig {
+    forbiddenTopics?: string[];
+    onViolation?: 'block' | 'flag';
+}
+export interface GroundingConfig {
+    enabled?: boolean;
+    context?: string[];
+    minOverlapRatio?: number;
+}
+export interface SecurityConfig {
+    preset?: SecurityPreset;
+    level?: SecurityLevel;
+    input?: {
+        maxContentLength?: number;
+        injectionDetection?: InjectionDetectionConfig;
+        pii?: PIIConfig;
+        secrets?: SecretDetectionConfig;
+        urls?: UrlRiskConfig;
+        tools?: ToolPolicyConfig;
+    };
+    output?: {
+        maxContentLength?: number;
+        piiRedaction?: boolean;
+        moderation?: ModerationConfig;
+        dlp?: DLPConfig;
+        topics?: TopicGuardrailConfig;
+        grounding?: GroundingConfig;
+    };
+}
+export interface SecurityFinding {
+    type: 'schema' | 'prompt-injection' | 'pii' | 'content-length' | 'secret' | 'url-risk' | 'tool-policy' | 'moderation' | 'dlp' | 'topic' | 'grounding';
+    severity: 'low' | 'medium' | 'high' | 'critical';
+    message: string;
+    path?: string;
+    value?: string;
+    metadata?: Record<string, unknown>;
+}
+export interface SecurityResult<T> {
+    ok: boolean;
+    value: T;
+    findings: SecurityFinding[];
+    guardrailsApplied: string[];
+}

package/dist/types/security.js

@@ -0,0 +1 @@
+export {};

package/dist/utils/ids.d.ts

@@ -0,0 +1 @@
+export declare function generateRequestId(): string;

package/dist/utils/ids.js

@@ -0,0 +1,4 @@
+import { randomBytes } from 'crypto';
+export function generateRequestId() {
+    return `req_${randomBytes(12).toString('hex')}`;
+}

package/dist/utils/logger.d.ts

@@ -0,0 +1,7 @@
+export declare class Logger {
+    private debug;
+    constructor(debug?: boolean);
+    info(msg: string, data?: Record<string, unknown>): void;
+    warn(msg: string, data?: Record<string, unknown>): void;
+    error(msg: string, err?: unknown): void;
+}

package/dist/utils/logger.js

@@ -0,0 +1,17 @@
+export class Logger {
+    debug;
+    constructor(debug = false) {
+        this.debug = debug;
+    }
+    info(msg, data) {
+        if (this.debug) {
+            console.log(`[nexus-ai-pro] ${msg}`, data ? JSON.stringify(data) : '');
+        }
+    }
+    warn(msg, data) {
+        console.warn(`[nexus-ai-pro] WARN: ${msg}`, data ? JSON.stringify(data) : '');
+    }
+    error(msg, err) {
+        console.error(`[nexus-ai-pro] ERROR: ${msg}`, err instanceof Error ? err.message : err);
+    }
+}

package/dist/utils/tokenizer.d.ts

@@ -0,0 +1,6 @@
+import type { CompletionRequest, Message } from '../types/messages.js';
+export declare class Tokenizer {
+    estimateTextTokens(text: string): number;
+    estimateMessageTokens(message: Message): number;
+    estimateRequestTokens(request: CompletionRequest): number;
+}