@ailib-official/ai-protocol 0.8.4

package/v1/models/hunyuan.yaml

@@ -0,0 +1,164 @@
+$schema: "https://raw.githubusercontent.com/ailib-official/ai-protocol/main/schemas/v1.json"
+protocol_version: "1.1"
+
+# 腾讯混元 (Tencent Hunyuan) Model Instances
+#
+# VERIFIED_AT: 2026-01-31 (based on public Tencent Cloud documentation)
+#
+# Notes:
+# - Hunyuan (混元) is Tencent's flagship LLM series.
+# - Strong performance in Chinese language tasks.
+
+models:
+  # Hunyuan Pro (flagship)
+  hunyuan-pro:
+    provider: hunyuan
+    model_id: "hunyuan-pro"
+    display_name: "混元 Pro"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+      notes: "Hunyuan flagship model with best quality."
+    context_window: 32768
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["tencent", "hunyuan", "pro", "flagship", "cn"]
+    pricing:
+      input_per_token: 0.00003
+      output_per_token: 0.0001
+
+  # Hunyuan Standard
+  hunyuan-standard:
+    provider: hunyuan
+    model_id: "hunyuan-standard"
+    display_name: "混元 Standard"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 32768
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["tencent", "hunyuan", "standard", "cn"]
+    pricing:
+      input_per_token: 0.0000045
+      output_per_token: 0.000005
+
+  hunyuan-standard-256k:
+    provider: hunyuan
+    model_id: "hunyuan-standard-256K"
+    display_name: "混元 Standard 256K"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 262144
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["tencent", "hunyuan", "standard", "ultra_long_context", "cn"]
+    pricing:
+      input_per_token: 0.000015
+      output_per_token: 0.00006
+
+  # Hunyuan Lite (cost-effective)
+  hunyuan-lite:
+    provider: hunyuan
+    model_id: "hunyuan-lite"
+    display_name: "混元 Lite"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+      notes: "Free tier available with rate limits."
+    context_window: 4096
+    capabilities: [chat, streaming]
+    status: active
+    tags: ["tencent", "hunyuan", "lite", "free_tier", "cn"]
+    pricing:
+      input_per_token: 0
+      output_per_token: 0
+
+  # Hunyuan Turbo (balanced speed and quality)
+  hunyuan-turbo:
+    provider: hunyuan
+    model_id: "hunyuan-turbo"
+    display_name: "混元 Turbo"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 32768
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["tencent", "hunyuan", "turbo", "cn"]
+    pricing:
+      input_per_token: 0.000015
+      output_per_token: 0.00005
+
+  hunyuan-turbo-latest:
+    provider: hunyuan
+    model_id: "hunyuan-turbo-latest"
+    display_name: "混元 Turbo Latest"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 32768
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["tencent", "hunyuan", "turbo", "latest", "cn"]
+    pricing:
+      input_per_token: 0.000015
+      output_per_token: 0.00005
+
+  # Hunyuan Vision
+  hunyuan-vision:
+    provider: hunyuan
+    model_id: "hunyuan-vision"
+    display_name: "混元 Vision"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 8192
+    capabilities: [chat, vision, streaming]
+    status: active
+    tags: ["tencent", "hunyuan", "vision", "multimodal", "cn"]
+    pricing:
+      input_per_token: 0.00002
+      output_per_token: 0.00002
+
+  # Hunyuan Code
+  hunyuan-code:
+    provider: hunyuan
+    model_id: "hunyuan-code"
+    display_name: "混元 Code"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 8192
+    capabilities: [chat, code, streaming]
+    status: active
+    tags: ["tencent", "hunyuan", "code", "cn"]
+    pricing:
+      input_per_token: 0.0000045
+      output_per_token: 0.000005
+
+  # Hunyuan Role (角色扮演)
+  hunyuan-role:
+    provider: hunyuan
+    model_id: "hunyuan-role"
+    display_name: "混元 Role"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 32768
+    capabilities: [chat, streaming]
+    status: active
+    tags: ["tencent", "hunyuan", "role", "roleplay", "cn"]
+    pricing:
+      input_per_token: 0.0000045
+      output_per_token: 0.000005

package/v1/models/jina.yaml

@@ -0,0 +1,98 @@
+$schema: "https://raw.githubusercontent.com/ailib-official/ai-protocol/main/schemas/v1.json"
+protocol_version: "1.5"
+
+# Jina AI Model Instances
+# Notes: Jina AI provides embeddings, rerank, and reader APIs (not chat LLM)
+# These models are primarily for RAG (Retrieval-Augmented Generation) infrastructure
+
+models:
+  # Text Embeddings
+  jina-embeddings-v3:
+    provider: jina
+    model_id: "jina-embeddings-v3"
+    display_name: "Jina Embeddings v3"
+    verification:
+      status: unverified
+      source: "manual"
+      notes: "State-of-the-art text embeddings"
+    context_window: 8192
+    capabilities: [embeddings]
+    modality: "text"
+    pricing:
+      input_per_token: null
+      output_per_token: null
+    status: active
+    tags: ["jina", "embeddings", "rag", "search"]
+    dimensions: 1024
+
+  jina-embeddings-v2-base-en:
+    provider: jina
+    model_id: "jina-embeddings-v2-base-en"
+    display_name: "Jina Embeddings v2 Base EN"
+    verification:
+      status: unverified
+      source: "manual"
+      notes: "English-only embeddings model"
+    context_window: 8192
+    capabilities: [embeddings]
+    modality: "text"
+    pricing:
+      input_per_token: null
+      output_per_token: null
+    status: active
+    tags: ["jina", "embeddings", "en", "fast"]
+    dimensions: 768
+
+  # Code Embeddings
+  jina-embeddings-v2-base-code:
+    provider: jina
+    model_id: "jina-embeddings-v2-base-code"
+    display_name: "Jina Embeddings v2 Base Code"
+    verification:
+      status: unverified
+      source: "manual"
+      notes: "Code embeddings for semantic code search"
+    context_window: 8192
+    capabilities: [embeddings]
+    modality: "code"
+    pricing:
+      input_per_token: null
+      output_per_token: null
+    status: active
+    tags: ["jina", "embeddings", "code", "search"]
+    dimensions: 768
+
+  # Rerank Models
+  jina-reranker-v2-base-en:
+    provider: jina
+    model_id: "jina-reranker-v2-base-en"
+    display_name: "Jina Reranker v2 Base EN"
+    verification:
+      status: unverified
+      source: "manual"
+      notes: "Re-ranking model for RAG"
+    context_window: 512
+    capabilities: [rerank]
+    modality: "text"
+    pricing:
+      input_per_token: null
+      output_per_token: null
+    status: active
+    tags: ["jina", "rerank", "rag", "search"]
+
+  jina-colbert-v2:
+    provider: jina
+    model_id: "jina-colbert-v2"
+    display_name: "Jina ColBERT v2"
+    verification:
+      status: unverified
+      source: "manual"
+      notes: "Late interaction reranking"
+    context_window: 512
+    capabilities: [rerank]
+    modality: "text"
+    pricing:
+      input_per_token: null
+      output_per_token: null
+    status: active
+    tags: ["jina", "rerank", "colbert", "rag"]

package/v1/models/lepton.yaml

@@ -0,0 +1,130 @@
+$schema: "https://raw.githubusercontent.com/ailib-official/ai-protocol/main/schemas/v1.json"
+protocol_version: "1.1"
+
+# Lepton AI Model Instances
+#
+# VERIFIED_AT: 2026-01-29 (based on public Lepton AI documentation)
+#
+# Notes:
+# - Lepton AI provides serverless inference for various models.
+# - Each model has its own subdomain endpoint.
+# - Base URL pattern: https://{model_id}.lepton.run/api/v1
+
+models:
+  # Llama 3.1 models
+  llama3-1-8b:
+    provider: lepton
+    model_id: "llama3-1-8b"
+    display_name: "Llama 3.1 8B"
+    verification:
+      status: unverified
+      verified_at: "2026-01-29"
+      source: "documentation"
+      notes: "Endpoint: https://llama3-1-8b.lepton.run/api/v1"
+    context_window: 131072
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["meta", "llama", "lepton"]
+
+  llama3-1-70b:
+    provider: lepton
+    model_id: "llama3-1-70b"
+    display_name: "Llama 3.1 70B"
+    verification:
+      status: unverified
+      verified_at: "2026-01-29"
+      source: "documentation"
+      notes: "Endpoint: https://llama3-1-70b.lepton.run/api/v1"
+    context_window: 131072
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["meta", "llama", "lepton"]
+
+  llama3-1-405b:
+    provider: lepton
+    model_id: "llama3-1-405b"
+    display_name: "Llama 3.1 405B"
+    verification:
+      status: unverified
+      verified_at: "2026-01-29"
+      source: "documentation"
+      notes: "Endpoint: https://llama3-1-405b.lepton.run/api/v1"
+    context_window: 131072
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["meta", "llama", "lepton", "flagship"]
+
+  # Mixtral models
+  mixtral-8x7b:
+    provider: lepton
+    model_id: "mixtral-8x7b"
+    display_name: "Mixtral 8x7B"
+    verification:
+      status: unverified
+      verified_at: "2026-01-29"
+      source: "documentation"
+      notes: "Endpoint: https://mixtral-8x7b.lepton.run/api/v1"
+    context_window: 32768
+    capabilities: [chat, streaming]
+    status: active
+    tags: ["mistral", "mixtral", "moe", "lepton"]
+
+  # Qwen models
+  qwen2-72b:
+    provider: lepton
+    model_id: "qwen2-72b"
+    display_name: "Qwen 2 72B"
+    verification:
+      status: unverified
+      verified_at: "2026-01-29"
+      source: "documentation"
+      notes: "Endpoint: https://qwen2-72b.lepton.run/api/v1"
+    context_window: 32768
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["alibaba", "qwen", "lepton"]
+
+  # Mistral models
+  mistral-7b:
+    provider: lepton
+    model_id: "mistral-7b"
+    display_name: "Mistral 7B"
+    verification:
+      status: unverified
+      verified_at: "2026-01-29"
+      source: "documentation"
+      notes: "Endpoint: https://mistral-7b.lepton.run/api/v1"
+    context_window: 8192
+    capabilities: [chat, streaming]
+    status: active
+    tags: ["mistral", "lepton"]
+
+  # Wizardlm models
+  wizardlm-2-8x22b:
+    provider: lepton
+    model_id: "wizardlm-2-8x22b"
+    display_name: "WizardLM 2 8x22B"
+    verification:
+      status: unverified
+      verified_at: "2026-01-29"
+      source: "documentation"
+      notes: "Endpoint: https://wizardlm-2-8x22b.lepton.run/api/v1"
+    context_window: 65536
+    capabilities: [chat, streaming]
+    status: active
+    tags: ["wizardlm", "moe", "lepton"]
+
+  # Dolphin models (uncensored)
+  dolphin-mixtral-8x7b:
+    provider: lepton
+    model_id: "dolphin-mixtral-8x7b"
+    display_name: "Dolphin Mixtral 8x7B"
+    verification:
+      status: unverified
+      verified_at: "2026-01-29"
+      source: "documentation"
+      notes: "Endpoint: https://dolphin-mixtral-8x7b.lepton.run/api/v1"
+    context_window: 32768
+    capabilities: [chat, streaming]
+    status: active
+    tags: ["dolphin", "mixtral", "lepton", "uncensored"]

package/v1/models/llama.yaml

@@ -0,0 +1,91 @@
+$schema: "https://raw.githubusercontent.com/ailib-official/ai-protocol/main/schemas/v1.json"
+protocol_version: "1.1"
+
+# Groq Llama Model Instances
+#
+# VERIFIED_AT: 2026-01-28 (via Groq list_models: GET /openai/v1/models)
+#
+# Notes:
+# - Groq model ids evolve over time; keep deprecated entries for historical reference.
+models:
+  # Historical ids (no longer present in current /models list for this key)
+  llama3-8b-8192:
+    provider: groq
+    model_id: "llama3-8b-8192"
+    display_name: "Llama 3 8B (legacy id)"
+    verification:
+      status: verified
+      verified_at: "2026-01-28"
+      source: "list_models:/models"
+      notes: "Legacy id (not present in current /models list); kept for reference."
+    context_window: 8192
+    capabilities: [chat, tools, streaming]
+    status: deprecated
+    tags: ["meta", "llama", "groq", "legacy"]
+
+  llama3-70b-8192:
+    provider: groq
+    model_id: "llama3-70b-8192"
+    display_name: "Llama 3 70B (legacy id)"
+    verification:
+      status: verified
+      verified_at: "2026-01-28"
+      source: "list_models:/models"
+      notes: "Legacy id (not present in current /models list); kept for reference."
+    context_window: 8192
+    capabilities: [chat, tools, streaming]
+    status: deprecated
+    tags: ["meta", "llama", "groq", "legacy"]
+
+  # Current Groq-available Llama family (examples)
+  llama-3.1-8b-instant:
+    provider: groq
+    model_id: "llama-3.1-8b-instant"
+    display_name: "Llama 3.1 8B Instant"
+    verification:
+      status: verified
+      verified_at: "2026-01-28"
+      source: "list_models:/models"
+    context_window: 8192
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["meta", "llama", "fast", "groq"]
+
+  llama-3.3-70b-versatile:
+    provider: groq
+    model_id: "llama-3.3-70b-versatile"
+    display_name: "Llama 3.3 70B Versatile"
+    verification:
+      status: verified
+      verified_at: "2026-01-28"
+      source: "list_models:/models"
+    context_window: 8192
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["meta", "llama", "groq"]
+
+  "meta-llama/llama-4-maverick-17b-128e-instruct":
+    provider: groq
+    model_id: "meta-llama/llama-4-maverick-17b-128e-instruct"
+    display_name: "Llama 4 Maverick 17B (128e) Instruct"
+    verification:
+      status: verified
+      verified_at: "2026-01-28"
+      source: "list_models:/models"
+    context_window: 8192
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["meta", "llama", "groq"]
+
+  "meta-llama/llama-4-scout-17b-16e-instruct":
+    provider: groq
+    model_id: "meta-llama/llama-4-scout-17b-16e-instruct"
+    display_name: "Llama 4 Scout 17B (16e) Instruct"
+    verification:
+      status: verified
+      verified_at: "2026-01-28"
+      source: "list_models:/models"
+    context_window: 8192
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["meta", "llama", "groq"]

package/v1/models/minimax.yaml

@@ -0,0 +1,132 @@
+$schema: "https://raw.githubusercontent.com/ailib-official/ai-protocol/main/schemas/v1.json"
+protocol_version: "1.1"
+
+# MiniMax Model Instances
+#
+# VERIFIED_AT: 2026-01-31 (based on public MiniMax platform documentation)
+#
+# Notes:
+# - MiniMax is a leading Chinese AI company.
+# - abab series are their flagship LLM models.
+# - Official docs: https://platform.minimaxi.com/document/ChatCompletion%20v2
+
+models:
+  # abab 6.5 series (latest flagship)
+  abab6.5s-chat:
+    provider: minimax
+    model_id: "abab6.5s-chat"
+    display_name: "MiniMax abab 6.5s Chat"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+      notes: "Latest flagship model with enhanced capabilities."
+    context_window: 245760
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["minimax", "abab", "flagship", "long_context", "cn"]
+    pricing:
+      input_per_token: 0.000001
+      output_per_token: 0.000001
+
+  abab6.5g-chat:
+    provider: minimax
+    model_id: "abab6.5g-chat"
+    display_name: "MiniMax abab 6.5g Chat"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+      notes: "General purpose flagship model."
+    context_window: 8192
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["minimax", "abab", "flagship", "cn"]
+    pricing:
+      input_per_token: 0.000001
+      output_per_token: 0.000001
+
+  abab6.5t-chat:
+    provider: minimax
+    model_id: "abab6.5t-chat"
+    display_name: "MiniMax abab 6.5t Chat"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+      notes: "Optimized for specific tasks."
+    context_window: 8192
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["minimax", "abab", "cn"]
+    pricing:
+      input_per_token: 0.000001
+      output_per_token: 0.000001
+
+  # abab 5.5 series (stable)
+  abab5.5s-chat:
+    provider: minimax
+    model_id: "abab5.5s-chat"
+    display_name: "MiniMax abab 5.5s Chat"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 16384
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["minimax", "abab", "stable", "cn"]
+    pricing:
+      input_per_token: 0.000001
+      output_per_token: 0.000001
+
+  abab5.5-chat:
+    provider: minimax
+    model_id: "abab5.5-chat"
+    display_name: "MiniMax abab 5.5 Chat"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 16384
+    capabilities: [chat, tools, streaming]
+    status: active
+    tags: ["minimax", "abab", "stable", "cn"]
+    pricing:
+      input_per_token: 0.0000005
+      output_per_token: 0.0000005
+
+  # Embedding model
+  embo-01:
+    provider: minimax
+    model_id: "embo-01"
+    display_name: "MiniMax Embedding"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+    context_window: 4096
+    capabilities: [embedding]
+    status: active
+    tags: ["minimax", "embedding", "cn"]
+    pricing:
+      input_per_token: 0.0000001
+      output_per_token: 0
+
+  # Speech model (T2A - Text to Audio)
+  speech-01-turbo:
+    provider: minimax
+    model_id: "speech-01-turbo"
+    display_name: "MiniMax Speech Turbo"
+    verification:
+      status: unverified
+      verified_at: "2026-01-31"
+      source: "documentation"
+      notes: "Text-to-speech model with multiple voice options."
+    context_window: 10000
+    capabilities: [speech_synthesis]
+    status: active
+    tags: ["minimax", "speech", "tts", "cn"]
+    pricing:
+      input_per_token: 0.000001
+      output_per_token: 0