Creating a Component
Component Specification
A Component Specification defines the metadata, container configuration, protocols, and behavioral contracts of a computational unit (component) that can be onboarded into the system. These components are indexed, made searchable and discoverable in the components registry, these components can be readily instantiated as blocks for serving.
This spec is processed by the add_component_IR function and saved into the Component collection in the database. Components can later be referenced in blocks, vDAGs, or management operations.
Top-Level Structure
The spec must be submitted under:
{
"body": {
"spec": {
"values": {
...
}
}
}
}
Fields
| Field | Type | Required | Description |
|---|---|---|---|
componentId |
object |
Yes | Structured identifier for the component. See breakdown below. |
componentId.name |
string |
Yes | Component name. |
componentId.version |
string |
Yes | Version of the component. |
componentId.releaseTag |
string |
Yes | Release tag for the component version. |
componentURI |
string |
Yes | Unique URI of the component. System-generated in most cases: <componentType>.<name>:<version>-<releaseTag> |
componentType |
string |
Yes | Type/category of the component (e.g., model, service, utility). |
containerRegistryInfo |
object |
No | Metadata about the container image associated with the component. See below. |
containerRegistryInfo.containerImage |
string |
Yes (if present) | Full container image path. |
containerRegistryInfo.containerRegistryId |
string |
Yes (if present) | Identifier of the registry the image is stored in. |
containerRegistryInfo.containerImageMetadata |
object |
No | Metadata like description, author, etc. |
containerRegistryInfo.componentMode |
string |
Yes (if present) | Enum: "aios" or "third_party" indicating the component’s origin. |
containerRegistryInfo.initContainer |
object |
Yes (if present) | Data of the init container if init container is specified, it's structure will be same as containerRegistryInfo |
componentMetadata |
object |
No | Custom metadata about the component. |
componentInitData |
object |
No | Initialization data or configuration required during component startup. |
componentInputProtocol |
object |
No | Description or schema of accepted inputs. |
componentOutputProtocol |
object |
No | Description or schema of generated outputs. |
policies |
object |
No | Policies attached to the component (e.g., lifecycle, security, governance). |
componentManagementCommandsTemplate |
object |
No | Defines supported management commands and their structures. |
componentInitSettings |
object |
No | Runtime settings such as flags, modes, or tuning parameters. |
componentParameters |
object |
No | Parameters used for processing/inference, typically at runtime. |
componentInitSettingsProtocol |
object |
No | Describes the expected format and structure of init settings. |
componentInitParametersProtocol |
object |
No | Describes the expected structure of runtime/inference parameters. |
tags |
array |
No | List of searchable tags associated with the component. |
Protocol Template Structure
Protocol templates define the structure, type, constraints, and description of data expected in a component's configuration, input, output, or parameters. These templates are JSON schemas that allow both human and machine interpretation of data contracts.
They are used in the following fields of a component specification:
componentInitSettingsProtocolcomponentInitParametersProtocolcomponentInputProtocolcomponentOutputProtocol
Each field within the template schema is defined with attributes such as data type, allowed values, validation rules, and optional nested structures.
Template Field Attributes
| Attribute | Type | Description |
|---|---|---|
type |
string |
The data type of the field. Allowed values: string, number, boolean, array, object, any. |
description |
string |
Human-readable description of the field’s purpose. |
pattern |
string |
(For strings) Regular expression the value must match. |
length |
number |
Maximum allowed string length. |
min, max |
number |
Minimum and maximum numeric values allowed. |
choices |
array |
List of valid values (enumeration). |
max_length |
number |
Maximum number of elements for arrays. |
properties |
object |
(For objects) Defines nested structure of key-value fields. |
items |
object |
(For arrays) Defines schema for each element of the array. |
Examples
componentInitSettingsProtocol
{
"logging_level": {
"type": "string",
"description": "Logging verbosity level",
"choices": ["debug", "info", "warn", "error"]
},
"enable_cache": {
"type": "boolean",
"description": "Whether to enable cache during execution"
},
"max_threads": {
"type": "number",
"description": "Maximum threads the component can spawn",
"min": 1,
"max": 64
}
}
componentInputProtocol
{
"image": {
"type": "string",
"description": "Base64 encoded image input",
"pattern": "^data:image/.*;base64,"
},
"metadata": {
"type": "object",
"description": "Optional metadata associated with the input",
"properties": {
"source": { "type": "string" },
"timestamp": { "type": "number" }
}
}
}
componentOutputProtocol
{
"predictions": {
"type": "array",
"description": "Top prediction results",
"max_length": 5,
"items": {
"type": "object",
"properties": {
"label": { "type": "string" },
"confidence": { "type": "number", "min": 0.0, "max": 1.0 }
}
}
},
"runtime_ms": {
"type": "number",
"description": "Inference time in milliseconds"
}
}
componentInitParametersProtocol
{
"threshold": {
"type": "number",
"description": "Minimum confidence required to accept a prediction",
"min": 0.0,
"max": 1.0
},
"top_k": {
"type": "number",
"description": "Number of top predictions to return",
"min": 1,
"max": 10
},
"class_filter": {
"type": "array",
"description": "Limit predictions to specific class labels",
"choices": ["cat", "dog", "car", "tree"],
"max_length": 4
}
}
Component Spec example: Object Detection
{
"body": {
"spec": {
"values": {
"componentId": {
"name": "object-detector",
"version": "2.0.0",
"releaseTag": "stable"
},
"componentURI": "model.object-detector:2.0.0-stable",
"componentType": "model",
"containerRegistryInfo": {
"containerImage": "registry.ai-platform.com/models/object-detector:2.0.0",
"containerRegistryId": "ai-platform-registry",
"containerImageMetadata": {
"author": "vision-team",
"description": "YOLO-based real-time object detection model"
},
"componentMode": "aios"
},
"componentMetadata": {
"usecase": "real-time object detection",
"framework": "PyTorch",
"hardware": "GPU"
},
"componentInitData": {
"weights_path": "/models/yolov5s.pt",
"device": "cuda"
},
"componentInputProtocol": {
"image": {
"type": "string",
"description": "Input image encoded as Base64",
"pattern": "^data:image/.*;base64,"
},
"image_shape": {
"type": "array",
"description": "Height and width of the input image",
"max_length": 2,
"items": {
"type": "number",
"min": 1
}
}
},
"componentOutputProtocol": {
"detections": {
"type": "array",
"description": "Detected objects",
"items": {
"type": "object",
"properties": {
"class": { "type": "string" },
"confidence": { "type": "number", "min": 0.0, "max": 1.0 },
"bbox": {
"type": "array",
"description": "Bounding box [x_min, y_min, x_max, y_max]",
"max_length": 4,
"items": { "type": "number" }
}
}
}
},
"processing_time_ms": {
"type": "number",
"description": "Inference time in milliseconds"
}
},
"componentInitParametersProtocol": {
"threshold": {
"type": "number",
"description": "Confidence threshold for filtering predictions",
"min": 0.0,
"max": 1.0
},
"top_k": {
"type": "number",
"description": "Return top-K detections per image",
"min": 1,
"max": 50
}
},
"componentInitSettingsProtocol": {
"enable_tracing": {
"type": "boolean",
"description": "Enable performance tracing logs"
},
"batch_size": {
"type": "number",
"description": "Batch size to be used during inference",
"min": 1,
"max": 32
}
},
"policies": {
"resource_affinity": {
"nodeType": "gpu",
"minMemory": "4GB"
}
},
"componentManagementCommandsTemplate": {
"restart": {
"description": "Restart the model service",
"args": {}
},
"reload_weights": {
"description": "Reload model weights from disk",
"args": {
"weights_path": {
"type": "string",
"required": true
}
}
}
},
"componentParameters": {
"threshold": 0.4,
"top_k": 10
},
"componentInitSettings": {
"enable_tracing": true,
"batch_size": 4
},
"tags": ["vision", "object-detection", "yolo", "realtime"]
}
}
}
}
Component spec example with init container:
{
"header": {
"templateUri": "Parser/V1",
"parameters": {}
},
"body": {
"spec": {
"values": {
"componentId": {
"name": "vllm-chat",
"version": "1.0.0",
"releaseTag": "stable"
},
"componentType": "model",
"containerRegistryInfo": {
"containerImage": "registry.ai-platform.com/models/vllm-external-chat:2.0.0",
"containerRegistryId": "ai-platform-registry",
"containerImageMetadata": {
"author": "llm-team",
"description": "LLM team"
},
"componentMode": "third-party",
"initContainer": {
"containerImage": "registry.ai-platform.com/init-containers/vLLM-deployer:2.0.0",
"containerRegistryId": "ai-platform-registry",
"containerImageMetadata": {
"author": "vision-team",
"description": "Init container to deploy vLLM system"
}
}
},
"componentMetadata": {
"usecase": "chat-completion",
"framework": "external-api",
"hardware": "external"
},
"componentInitData": {
"vllm_server_url": "http://vllm-service:8080",
"model_name": "mistralai/Mistral-7B-Instruct-v0.1"
},
"componentInitParametersProtocol": {
"temperature": {
"type": "number",
"description": "Sampling temperature",
"min": 0.0,
"max": 2.0
}
},
"componentInitSettingsProtocol": {
"max_new_tokens": {
"type": "number",
"description": "Maximum number of tokens to generate",
"min": 1,
"max": 2048
}
},
"componentInputProtocol": {
"message": {
"type": "string",
"description": "Input message from the user"
},
"session_id": {
"type": "string",
"description": "Unique identifier for the chat session"
}
},
"componentOutputProtocol": {
"reply": {
"type": "string",
"description": "Chat reply from the vLLM model"
}
},
"componentParameters": {
"temperature": 0.7
},
"componentInitSettings": {
"max_new_tokens": 256
},
"componentManagementCommandsTemplate": {
"reset": {
"description": "Clears all stored chat sessions",
"args": {}
}
},
"tags": [
"vllm",
"chat",
"openai-compatible",
"mistral",
"llm"
]
}
}
}
}
Onboarding the component:
curl -X POST http://<server-url>/api/addComponent \
-H "Content-Type: application/json" \
-d @component_spec.json