@ailib-official/ai-protocol 0.8.4

package/dist/v1/models/zhipu.json

@@ -0,0 +1,250 @@
+{
+  "$schema": "https://raw.githubusercontent.com/ailib-official/ai-protocol/main/schemas/v1.json",
+  "protocol_version": "1.1",
+  "models": {
+    "glm-4-plus": {
+      "provider": "zhipu",
+      "model_id": "glm-4-plus",
+      "display_name": "GLM-4 Plus",
+      "verification": {
+        "status": "unverified",
+        "verified_at": "2026-01-31",
+        "source": "documentation",
+        "notes": "Latest flagship model with best performance."
+      },
+      "context_window": 131072,
+      "capabilities": [
+        "chat",
+        "tools",
+        "streaming"
+      ],
+      "status": "active",
+      "tags": [
+        "zhipu",
+        "glm",
+        "flagship",
+        "cn"
+      ],
+      "pricing": {
+        "input_per_token": 0.00005,
+        "output_per_token": 0.00005
+      }
+    },
+    "glm-4-0520": {
+      "provider": "zhipu",
+      "model_id": "glm-4-0520",
+      "display_name": "GLM-4 0520",
+      "verification": {
+        "status": "unverified",
+        "verified_at": "2026-01-31",
+        "source": "documentation"
+      },
+      "context_window": 131072,
+      "capabilities": [
+        "chat",
+        "tools",
+        "streaming"
+      ],
+      "status": "active",
+      "tags": [
+        "zhipu",
+        "glm",
+        "cn"
+      ],
+      "pricing": {
+        "input_per_token": 0.0001,
+        "output_per_token": 0.0001
+      }
+    },
+    "glm-4-air": {
+      "provider": "zhipu",
+      "model_id": "glm-4-air",
+      "display_name": "GLM-4 Air",
+      "verification": {
+        "status": "unverified",
+        "verified_at": "2026-01-31",
+        "source": "documentation",
+        "notes": "Cost-effective model with good performance."
+      },
+      "context_window": 131072,
+      "capabilities": [
+        "chat",
+        "tools",
+        "streaming"
+      ],
+      "status": "active",
+      "tags": [
+        "zhipu",
+        "glm",
+        "cost_effective",
+        "cn"
+      ],
+      "pricing": {
+        "input_per_token": 0.000001,
+        "output_per_token": 0.000001
+      }
+    },
+    "glm-4-airx": {
+      "provider": "zhipu",
+      "model_id": "glm-4-airx",
+      "display_name": "GLM-4 AirX",
+      "verification": {
+        "status": "unverified",
+        "verified_at": "2026-01-31",
+        "source": "documentation",
+        "notes": "Fast inference version of GLM-4 Air."
+      },
+      "context_window": 8192,
+      "capabilities": [
+        "chat",
+        "tools",
+        "streaming"
+      ],
+      "status": "active",
+      "tags": [
+        "zhipu",
+        "glm",
+        "fast",
+        "cn"
+      ],
+      "pricing": {
+        "input_per_token": 0.00001,
+        "output_per_token": 0.00001
+      }
+    },
+    "glm-4-flash": {
+      "provider": "zhipu",
+      "model_id": "glm-4-flash",
+      "display_name": "GLM-4 Flash",
+      "verification": {
+        "status": "unverified",
+        "verified_at": "2026-01-31",
+        "source": "documentation",
+        "notes": "Free tier model with rate limits."
+      },
+      "context_window": 131072,
+      "capabilities": [
+        "chat",
+        "tools",
+        "streaming"
+      ],
+      "status": "active",
+      "tags": [
+        "zhipu",
+        "glm",
+        "free_tier",
+        "cn"
+      ],
+      "pricing": {
+        "input_per_token": 0,
+        "output_per_token": 0
+      }
+    },
+    "glm-4-long": {
+      "provider": "zhipu",
+      "model_id": "glm-4-long",
+      "display_name": "GLM-4 Long",
+      "verification": {
+        "status": "unverified",
+        "verified_at": "2026-01-31",
+        "source": "documentation"
+      },
+      "context_window": 1048576,
+      "capabilities": [
+        "chat",
+        "streaming"
+      ],
+      "status": "active",
+      "tags": [
+        "zhipu",
+        "glm",
+        "ultra_long_context",
+        "cn"
+      ],
+      "pricing": {
+        "input_per_token": 0.000001,
+        "output_per_token": 0.000001
+      }
+    },
+    "glm-4v-plus": {
+      "provider": "zhipu",
+      "model_id": "glm-4v-plus",
+      "display_name": "GLM-4V Plus",
+      "verification": {
+        "status": "unverified",
+        "verified_at": "2026-01-31",
+        "source": "documentation"
+      },
+      "context_window": 8192,
+      "capabilities": [
+        "chat",
+        "vision",
+        "streaming"
+      ],
+      "status": "active",
+      "tags": [
+        "zhipu",
+        "glm",
+        "vision",
+        "multimodal",
+        "cn"
+      ],
+      "pricing": {
+        "input_per_token": 0.00001,
+        "output_per_token": 0.00001
+      }
+    },
+    "glm-4v": {
+      "provider": "zhipu",
+      "model_id": "glm-4v",
+      "display_name": "GLM-4V",
+      "verification": {
+        "status": "unverified",
+        "verified_at": "2026-01-31",
+        "source": "documentation"
+      },
+      "context_window": 2048,
+      "capabilities": [
+        "chat",
+        "vision",
+        "streaming"
+      ],
+      "status": "active",
+      "tags": [
+        "zhipu",
+        "glm",
+        "vision",
+        "multimodal",
+        "cn"
+      ],
+      "pricing": {
+        "input_per_token": 0.00005,
+        "output_per_token": 0.00005
+      }
+    },
+    "embedding-3": {
+      "provider": "zhipu",
+      "model_id": "embedding-3",
+      "display_name": "Zhipu Embedding 3",
+      "verification": {
+        "status": "unverified",
+        "verified_at": "2026-01-31",
+        "source": "documentation"
+      },
+      "context_window": 8192,
+      "capabilities": [
+        "embedding"
+      ],
+      "status": "active",
+      "tags": [
+        "zhipu",
+        "embedding",
+        "cn"
+      ],
+      "pricing": {
+        "input_per_token": 5e-7,
+        "output_per_token": 0
+      }
+    }
+  }
+}

package/dist/v1/providers/ai21.json

@@ -0,0 +1,215 @@
+{
+  "$schema": "https://raw.githubusercontent.com/ailib-official/ai-protocol/main/schemas/v1.json",
+  "id": "ai21",
+  "protocol_version": "1.5",
+  "name": "AI21 Labs",
+  "version": "v1",
+  "status": "stable",
+  "category": "ai_provider",
+  "official_url": "https://docs.ai21.com",
+  "support_contact": "https://www.ai21.com/contact",
+  "endpoint": {
+    "base_url": "https://api.ai21.com/studio/v1",
+    "protocol": "https",
+    "timeout_ms": 30000
+  },
+  "auth": {
+    "type": "bearer",
+    "token_env": "AI21_API_KEY"
+  },
+  "payload_format": "ai21_style",
+  "api_families": [
+    "chat",
+    "complete"
+  ],
+  "default_api_family": "chat",
+  "endpoints": {
+    "chat": {
+      "path": "/chat/completions",
+      "method": "POST",
+      "adapter": "ai21_chat"
+    },
+    "jamba_chat": {
+      "path": "/jamba-1.5-large/chat/completions",
+      "method": "POST",
+      "adapter": "openai"
+    }
+  },
+  "services": {
+    "list_models": {
+      "path": "/models",
+      "method": "GET"
+    }
+  },
+  "termination": {
+    "source_field": "finish_reason",
+    "mapping": {
+      "stop": "end_turn",
+      "length": "max_tokens",
+      "endoftext": "end_turn"
+    },
+    "notes": [
+      "AI21 uses finish_reason similar to OpenAI.",
+      "Jamba models follow OpenAI-compatible format."
+    ]
+  },
+  "tooling": {
+    "source_model": "openai_tool_calls",
+    "tool_use": {
+      "id_path": "id",
+      "name_path": "function.name",
+      "input_path": "function.arguments",
+      "input_format": "json_string"
+    },
+    "notes": [
+      "Jamba models support function calling in OpenAI format."
+    ]
+  },
+  "retry_policy": {
+    "strategy": "exponential_backoff",
+    "max_retries": 3,
+    "min_delay_ms": 1000,
+    "max_delay_ms": 30000,
+    "jitter": "full",
+    "retry_on_http_status": [
+      429,
+      500,
+      502,
+      503
+    ],
+    "notes": [
+      "AI21 uses standard rate limiting."
+    ]
+  },
+  "error_classification": {
+    "by_http_status": {
+      "400": "invalid_request",
+      "401": "authentication",
+      "403": "permission_denied",
+      "404": "not_found",
+      "422": "invalid_request",
+      "429": "rate_limited",
+      "500": "server_error",
+      "503": "overloaded"
+    }
+  },
+  "parameter_mappings": {
+    "temperature": "temperature",
+    "max_tokens": "max_tokens",
+    "stream": "stream",
+    "top_p": "top_p",
+    "stop_sequences": "stop",
+    "tools": "tools",
+    "tool_choice": "tool_choice",
+    "n": "n"
+  },
+  "response_format": "ai21_style",
+  "response_paths": {
+    "content": "choices[0].message.content",
+    "tool_calls": "choices[0].message.tool_calls",
+    "usage": "usage",
+    "finish_reason": "choices[0].finish_reason"
+  },
+  "streaming": {
+    "event_format": "sse",
+    "decoder": {
+      "format": "sse",
+      "delimiter": "\n\n",
+      "prefix": "data: ",
+      "done_signal": "[DONE]"
+    },
+    "content_path": "choices[0].delta.content",
+    "tool_call_path": "choices[0].delta.tool_calls",
+    "usage_path": "usage",
+    "frame_selector": "exists($.choices) || exists($.error)",
+    "event_map": [
+      {
+        "match": "exists($.choices[*].delta.content)",
+        "emit": "PartialContentDelta",
+        "fields": {
+          "content": "$.choices[*].delta.content"
+        }
+      },
+      {
+        "match": "exists($.choices[*].delta.tool_calls[*].function.name)",
+        "emit": "ToolCallStarted",
+        "fields": {
+          "tool_call_id": "$.choices[*].delta.tool_calls[*].id",
+          "tool_name": "$.choices[*].delta.tool_calls[*].function.name",
+          "index": "$.choices[*].delta.tool_calls[*].index"
+        }
+      },
+      {
+        "match": "exists($.choices[*].delta.tool_calls[*].function.arguments)",
+        "emit": "PartialToolCall",
+        "fields": {
+          "arguments": "$.choices[*].delta.tool_calls[*].function.arguments",
+          "index": "$.choices[*].delta.tool_calls[*].index"
+        }
+      },
+      {
+        "match": "exists($.usage)",
+        "emit": "Metadata",
+        "fields": {
+          "usage": "$.usage"
+        }
+      },
+      {
+        "match": "exists($.choices[*].finish_reason)",
+        "emit": "FinalCandidate",
+        "fields": {
+          "finish_reason": "$.choices[*].finish_reason"
+        }
+      }
+    ],
+    "stop_condition": "$.choices[0].finish_reason != null"
+  },
+  "features": {
+    "multi_candidate": {
+      "support_type": "native",
+      "param_name": "n"
+    },
+    "response_mapping": {
+      "tool_calls": {
+        "path": "choices[0].message.tool_calls",
+        "fields": {
+          "id": "id",
+          "name": "function.name",
+          "args": "function.arguments"
+        }
+      },
+      "error": {
+        "message_path": "detail",
+        "type_path": "type"
+      }
+    }
+  },
+  "capabilities": {
+    "streaming": true,
+    "tools": true,
+    "vision": false,
+    "agentic": true,
+    "parallel_tools": false,
+    "reasoning": false
+  },
+  "availability": {
+    "required": false,
+    "regions": [
+      "global"
+    ],
+    "check": {
+      "method": "GET",
+      "path": "/models",
+      "expected_status": [
+        200,
+        401
+      ],
+      "timeout_ms": 5000
+    }
+  },
+  "experimental_features": [
+    "contextual_answers",
+    "summarization",
+    "paraphrase"
+  ]
+}