Models

`smarter.models.base`

The base classes of the SDK’s response models.

Every model allows extra fields. The Smarter platform adds fields to its responses in minor releases, and an SDK that rejected them would break whenever the platform is upgraded. Extra fields are kept, and are available as attributes and in SmarterModel.to_dict().

class smarter.models.base.SmarterModel

Bases: pydantic.main.BaseModel

The base class of the SDK’s models: a Pydantic model that keeps unknown fields.

to_dict() → dict[str, Any]

The model as a JSON-compatible dict, with the api’s own field names.

Returns:: The fields that the api returned, or that were set.

to_json(indent: int | None = 2) → str

The model as JSON text, with the api’s own field names.

Parameters:: indent – The indentation, or None for compact JSON.
Returns:: The JSON text.

class smarter.models.base.SmarterCamelModel

Bases: smarter.models.base.SmarterModel

A model of a response whose field names are camelCase, e.g. hashedId.

Its attributes are snake_case (hashed_id), and it accepts either spelling.

class smarter.models.base.SmarterUserRef

Bases: smarter.models.base.SmarterCamelModel

The user of a resource’s owner.

username: str

email: Optional

class smarter.models.base.SmarterAccountRef

Bases: smarter.models.base.SmarterCamelModel

The account of a resource’s owner.

account_number: str = alias ``accountNumber``

class smarter.models.base.SmarterOwner

Bases: smarter.models.base.SmarterCamelModel

The owner of a resource: its user profile, as the user and the account.

user: Optional

account: Optional

class smarter.models.base.SmarterResource

Bases: smarter.models.base.SmarterCamelModel

The fields that every Smarter resource shares, as the REST api returns them.

The api’s resources are Django rows with the same metadata as a manifest’s metadata: a name, description, version, tags and annotations. Each resource’s model adds its own fields.

id: int

hashed_id: Optional = alias ``hashedId``

name: str

description: Optional

version: Optional

tags: list

annotations: list

created_at: Optional = alias ``createdAt``

updated_at: Optional = alias ``updatedAt``

user_profile: Optional = alias ``userProfile``

can_delete: Optional = alias ``canDelete``

manifest_url: Optional = alias ``manifestUrl``

ready: Optional

`smarter.models.manifest`

Models of Smarter Application Manifests (SAM): the YAML documents that declare every Smarter resource.

A manifest has the same four sections as a Kubernetes manifest: apiVersion, kind, metadata and spec. describe adds a read-only status. The shape of spec depends on the kind and the platform’s version, so it is a JSON object here. The platform publishes the complete JSON Schema of each kind: see smarter.resources.manifests.Manifests.json_schema().

smarter.models.manifest.SMARTER_API_VERSION: The apiVersion of every Smarter manifest.

class smarter.models.manifest.ManifestMetadata

Bases: smarter.models.base.SmarterModel

A manifest’s metadata: the resource’s name, and descriptive information about it.

name: str: The resource’s name, unique per account and kind, in snake_case.

description: Optional

version: Optional: The resource’s semantic version, e.g. 1.0.0.

tags: list

annotations: list: Key-value pairs, as a list of single-key objects.

class smarter.models.manifest.Manifest

Bases: smarter.models.base.SmarterModel

A Smarter Application Manifest.

manifest = Manifest.from_yaml(open("llmclient.yaml").read())
manifest.metadata.name
manifest.spec["config"]["provider"]

apiVersion: str

kind: str: The resource kind, e.g. LLMClient. See Smarter.kinds().

metadata: ManifestMetadata

spec: dict: The resource’s specification.

status: Optional: Read-only state, returned by describe.

classmethod from_yaml(text: str) → 'Manifest'

Parse a manifest from YAML (or JSON) text.

Parameters:: text – The manifest’s text.
Returns:: The manifest.
Raises:: SmarterValueError – If the text is not a YAML object.

classmethod from_file(path: str | pathlib.Path) → 'Manifest'

Read a manifest from a YAML or JSON file.

Parameters:: path – The file’s path.
Returns:: The manifest.

to_sam() → 'SAMModel'

The manifest as its kind’s typed SAM model, e.g. a SAMGuardrail.

guardrail = client.guardrails.describe("pii_redactor").to_sam()
print(guardrail.spec.config.action)

Returns:

The typed manifest.

Raises:

SmarterValueError – If the SDK has no model of the manifest’s kind.
pydantic.ValidationError – If the manifest does not match its kind’s model.

to_yaml() → str

The manifest as YAML text, without its read-only status, ready to apply.

Returns:: The YAML text.

class smarter.models.manifest.ManifestListTitle

Bases: smarter.models.base.SmarterModel

A column of a get result: the field’s name, and its Django field type.

name: str

type: str

class smarter.models.manifest.ManifestListData

Bases: smarter.models.base.SmarterModel

The table of a get result.

titles: list

items: list

class smarter.models.manifest.ManifestListMetadata

Bases: smarter.models.base.SmarterModel

The metadata of a get result.

count: int

class smarter.models.manifest.ManifestList

Bases: smarter.models.base.SmarterModel

The result of get: the resources of one kind, as a table.

data.items are the rows, and data.titles their columns. The columns depend on the kind.

apiVersion: str

kind: str

name: Optional

metadata: ManifestListMetadata

data: ManifestListData

property items: The rows.

property names: The names of the resources: each row’s name, or its username for the User kind.

class smarter.models.manifest.KindInfo

Bases: smarter.models.base.SmarterCamelModel

A resource kind that the platform supports, as /api/v1/cli/resources/ describes it.

singular: str = alias ``Singular``

plural: str = alias ``Plural``

api_kind: str = alias ``APIKind``: The kind, as it appears in a manifest.

display: Optional = alias ``Display``

display_plural: Optional = alias ``DisplayPlural``

deployable: bool = alias ``Deployable``: True if the kind can be deployed.

`smarter.models.platform`

Models of the platform’s own endpoints: whoami, version and status.

class smarter.models.platform.WhoAmI

Bases: smarter.models.base.SmarterModel

The owner of the api key: the user, their account, and the platform’s environment.

user: User

account: Account

environment: Optional: The platform’s environment, e.g. local or prod.

class smarter.models.platform.ApiVersion

Bases: smarter.models.base.SmarterModel

The platform’s version, and the versions of its main Python packages.

version: str: The Smarter platform’s semantic version.

python: dict: Python package versions, by package.

class smarter.models.platform.CliVersion

Bases: smarter.models.base.SmarterModel

The version of the smarter cli that the platform recommends.

version: Optional

class smarter.models.platform.PlatformVersion

Bases: smarter.models.base.SmarterModel

The response of version.

api: ApiVersion

cli: Optional

class smarter.models.platform.PlatformStatus

Bases: smarter.models.base.SmarterCamelModel

The response of status: the state of the platform’s infrastructure and compute.

The sections’ contents depend on how the platform is hosted, so they are JSON objects.

infrastructures: dict

compute: dict

`smarter.models.account`

Models of accounts, users and account contacts, as /api/v1/accounts/ returns them.

class smarter.models.account.Account

Bases: smarter.models.base.SmarterResource

A Smarter account: the tenant that owns users and resources.

account_number: str = alias ``accountNumber``

is_default_account: Optional = alias ``isDefaultAccount``

company_name: Optional = alias ``companyName``

phone_number: Optional = alias ``phoneNumber``

address1: Optional

address2: Optional

city: Optional

state: Optional

postal_code: Optional = alias ``postalCode``

country: Optional

language: Optional

timezone: Optional

currency: Optional

is_active: Optional = alias ``isActive``

class smarter.models.account.User

Bases: smarter.models.base.SmarterCamelModel

A user of the platform.

id: int

username: str

first_name: Optional = alias ``firstName``

last_name: Optional = alias ``lastName``

email: Optional

is_staff: Optional = alias ``isStaff``

is_superuser: Optional = alias ``isSuperuser``

class smarter.models.account.AccountContact

Bases: smarter.models.base.SmarterCamelModel

A contact of an account, who receives its notices.

id: int

account: Optional

created_at: Optional = alias ``createdAt``

updated_at: Optional = alias ``updatedAt``

first_name: Optional = alias ``firstName``

last_name: Optional = alias ``lastName``

email: Optional

phone: Optional

is_primary: Optional = alias ``isPrimary``

is_test: Optional = alias ``isTest``

welcomed: Optional

`smarter.models.llmclient`

Models of LLMClients and their plugins, built-in functions, api keys and custom domains.

class smarter.models.llmclient.LLMClientPluginRef

Bases: smarter.models.base.SmarterCamelModel

A plugin of an LLMClient, by its id and name.

id: int

name: Optional

class smarter.models.llmclient.LLMClientFunction

Bases: smarter.models.base.SmarterCamelModel

A built-in function of an LLMClient, e.g. calculator.

id: int

name: Optional

llmclient: Optional

created_at: Optional = alias ``createdAt``

updated_at: Optional = alias ``updatedAt``

class smarter.models.llmclient.LLMClientPlugin

Bases: smarter.models.base.SmarterCamelModel

The association of a plugin with an LLMClient.

id: int

plugin_meta: Plugin = alias ``pluginMeta``

llmclient: Optional

class smarter.models.llmclient.LLMClientApiKey

Bases: smarter.models.base.SmarterCamelModel

An api key that may call an LLMClient that requires authentication.

id: int

llmclient: Optional

api_key: Optional = alias ``apiKey``

class smarter.models.llmclient.LLMClientCustomDomain

Bases: smarter.models.base.SmarterCamelModel

A custom domain on which an LLMClient is deployed.

id: int

llmclient: Optional

custom_domain: Optional = alias ``customDomain``

class smarter.models.llmclient.LLMClient

Bases: smarter.models.base.SmarterResource

An LLMClient: a deployable AI agent, built from an LLM provider and model, a system prompt, plugins and functions.

Prompt it with smarter.resources.llmclients.LLMClients.prompt(), or start a chat with smarter.resources.llmclients.LLMClients.chat().

rfc1034_compliant_name: Optional = alias ``rfc1034CompliantName``

provider: Optional

default_model: Optional = alias ``defaultModel``

default_system_role: Optional = alias ``defaultSystemRole``: The system prompt of a new chat session.

deployed: Optional

default_host: Optional = alias ``defaultHost``

default_url: Optional = alias ``defaultUrl``

custom_host: Optional = alias ``customHost``

custom_url: Optional = alias ``customUrl``

sandbox_host: Optional = alias ``sandboxHost``

sandbox_url: Optional = alias ``sandboxUrl``

functions: list

plugins: list

`smarter.models.plugin`

Models of plugins: the tools that an LLMClient’s LLM can call.

class smarter.models.plugin.Plugin

Bases: smarter.models.base.SmarterResource

A plugin’s metadata.

Its kind is the manifest kind, e.g. ApiPlugin or SqlPlugin, and its plugin_class the implementation: api, skill, sql, static, websearch or imagesearch.

plugin_class: Optional = alias ``pluginClass``: api, skill, sql, static, websearch or imagesearch.

kind: Optional: The manifest kind, e.g. ApiPlugin.

`smarter.models.guardrail`

Models of guardrails, their evaluations, and their events.

class smarter.models.guardrail.Guardrail

Bases: smarter.models.base.SmarterResource

A guardrail: a check that the platform runs on prompts (input) or on the LLM’s replies (output).

stage: Optional: input or output.

category: Optional

strategy: Optional: e.g. regex, detector, classifier or llm.

pattern: Optional

config: Optional

threshold: Optional

action: Optional: e.g. block, redact or flag.

replacement: Optional

message: Optional

severity: Optional

mode: Optional: enforce or monitor.

fail_closed: Optional = alias ``failClosed``

priority: Optional

is_active: Optional = alias ``isActive``

class smarter.models.guardrail.GuardrailMatch

Bases: smarter.models.base.SmarterModel

A span of the text that a guardrail matched.

start: int

end: int

text: str

label: Optional

class smarter.models.guardrail.GuardrailFinding

Bases: smarter.models.base.SmarterModel

One finding of a guardrail evaluation.

confidence: Optional

rationale: Optional

matches: list

class smarter.models.guardrail.GuardrailEvaluation

Bases: smarter.models.base.SmarterCamelModel

The result of a guardrail’s dry run on a text.

Nothing is saved: an evaluation does not create an event.

guardrail: str

stage: str

triggered: bool

disposition: Optional: e.g. passed, blocked, redacted or flagged.

text: Optional: The text after the guardrail’s action, e.g. redacted.

message: Optional

error: Optional

duration_ms: Optional = alias ``durationMs``

findings: list

class smarter.models.guardrail.GuardrailEvent

Bases: smarter.models.base.SmarterCamelModel

A guardrail’s finding in a real prompt, saved for review.

id: int

created_at: Optional = alias ``createdAt``

llmclient: Optional: The name of the LLMClient whose prompt triggered it.

stage: Optional

disposition: Optional

reviewed: Optional

class smarter.models.guardrail.GuardrailEvents

Bases: smarter.models.base.SmarterModel

The latest events of a guardrail.

guardrail: str

events: list

`smarter.models.mcpclient`

Models of MCPClients: connections to Model Context Protocol servers, whose tools an LLMClient’s LLM can call.

class smarter.models.mcpclient.MCPClient

Bases: smarter.models.base.SmarterResource

An MCPClient: an MCP server, how to reach it, and which of its tools the LLM may call.

transport: Optional: http, sse or stdio.

endpoint_url: Optional = alias ``endpointUrl``

command: Optional

timeout: Optional

auth_type: Optional = alias ``authType``

allowed_tools: list = alias ``allowedTools``

allowed_resources: list = alias ``allowedResources``

include_instructions: Optional = alias ``includeInstructions``

cache_ttl: Optional = alias ``cacheTtl``

is_active: Optional = alias ``isActive``

status: Optional: e.g. connected or error.

priority: Optional

protocol_version: Optional = alias ``protocolVersion``

server_name: Optional = alias ``serverName``

server_version: Optional = alias ``serverVersion``

tools: list: The names of every tool that the server offers.

last_connected_at: Optional = alias ``lastConnectedAt``

last_error: Optional = alias ``lastError``

class smarter.models.mcpclient.MCPClientStatus

Bases: smarter.models.base.SmarterCamelModel

The connection state of an MCPClient.

name: str

status: Optional

protocol_version: Optional = alias ``protocolVersion``

server_name: Optional = alias ``serverName``

server_version: Optional = alias ``serverVersion``

last_connected: Optional = alias ``lastConnected``

last_error: Optional = alias ``lastError``

error: Optional: The error of a failed refresh.

class smarter.models.mcpclient.MCPTool

Bases: smarter.models.base.SmarterCamelModel

A tool of an MCP server, which the MCPClient allows the LLM to call.

name: str

title: Optional

description: Optional

input_schema: Optional = alias ``inputSchema``: The tool’s arguments, as a JSON Schema.

read_only: Optional = alias ``readOnly``

destructive: Optional

class smarter.models.mcpclient.MCPClientTools

Bases: smarter.models.mcpclient.MCPClientStatus

An MCPClient’s allowed tools, and the server’s instructions.

instructions: Optional

tools: list

class smarter.models.mcpclient.MCPToolResult

Bases: smarter.models.base.SmarterCamelModel

The result of a tool call.

name: str: The MCPClient’s name.

tool: str

is_error: bool = alias ``isError``

content: Optional: The tool’s result, as text.

`smarter.models.vectorstore`

Models of vectorstores, their documents and snapshots, and search results.

class smarter.models.vectorstore.Vectorstore

Bases: smarter.models.base.SmarterResource

A vectorstore: a vector database, hosted by Smarter or by a provider, for retrieval-augmented generation.

backend: Optional: e.g. qdrant or pinecone.

hosting: Optional: self_hosted or managed.

is_active: Optional = alias ``isActive``

dimension: Optional

metric: Optional

embeddings_provider: Optional = alias ``embeddingsProvider``

embeddings_model: Optional = alias ``embeddingsModel``

status: Optional: e.g. pending, ready or error.

status_message: Optional = alias ``statusMessage``

index_name: Optional = alias ``indexName``

endpoint_url: Optional = alias ``endpointUrl``

vector_count: Optional = alias ``vectorCount``

document_count: Optional = alias ``documentCount``

snapshot_count: Optional = alias ``snapshotCount``

spec: Optional

deployed_at: Optional = alias ``deployedAt``

class smarter.models.vectorstore.VectorstoreStatus

Bases: smarter.models.base.SmarterCamelModel

A vectorstore’s state, checked now.

vectorstore: str

status: str

message: Optional

vector_count: Optional = alias ``vectorCount``

stats: dict

class smarter.models.vectorstore.VectorstoreDocument

Bases: smarter.models.base.SmarterCamelModel

A document of a vectorstore: a file, a text or a url, split into chunks and embedded.

id: int

name: str

created_at: Optional = alias ``createdAt``

updated_at: Optional = alias ``updatedAt``

source: Optional: upload, text or url.

content_type: Optional = alias ``contentType``

status: Optional: e.g. pending, loaded, error or deleting.

chunk_count: Optional = alias ``chunkCount``

metadata: Optional

class smarter.models.vectorstore.VectorstoreDocuments

Bases: smarter.models.base.SmarterCamelModel

A vectorstore’s documents.

documents: list

class smarter.models.vectorstore.VectorstoreDocumentsAdded

Bases: smarter.models.base.SmarterCamelModel

The result of adding documents.

Documents are loaded asynchronously, and only once the vectorstore is deployed: see queued.

documents: list

duplicates: list: Documents that already existed.

queued: bool

message: Optional

class smarter.models.vectorstore.VectorstoreDocumentDeleted

Bases: smarter.models.base.SmarterCamelModel

A document that is being deleted, with its chunks.

document: str

status: str

class smarter.models.vectorstore.VectorstoreSearchResult

Bases: smarter.models.base.SmarterCamelModel

A chunk that matched a search.

id: Optional

text: str

metadata: Optional

score: Optional

class smarter.models.vectorstore.VectorstoreSearchResults

Bases: smarter.models.base.SmarterCamelModel

The chunks that matched a search, best first.

vectorstore: str

results: list

class smarter.models.vectorstore.VectorstoreSnapshot

Bases: smarter.models.base.SmarterCamelModel

A snapshot of a vectorstore’s data.

id: int

name: Optional

created_at: Optional = alias ``createdAt``

status: Optional

scheduled: Optional

size_bytes: Optional = alias ``sizeBytes``

class smarter.models.vectorstore.VectorstoreSnapshots

Bases: smarter.models.base.SmarterCamelModel

A vectorstore’s snapshots.

snapshots: list

class smarter.models.vectorstore.VectorstoreRestored

Bases: smarter.models.base.SmarterCamelModel

The result of restoring a snapshot.

vectorstore: str

restored: str

vector_count: Optional = alias ``vectorCount``

`smarter.models.llmhost`

Models of LLMHosts: open-weight models that Smarter serves on its own Kubernetes cluster.

class smarter.models.llmhost.LLMHost

Bases: smarter.models.base.SmarterResource

An LLMHost: a model, the inference engine that serves it, and the compute it runs on.

spec: Optional

model_source: Optional = alias ``modelSource``

model_repository: Optional = alias ``modelRepository``

model_revision: Optional = alias ``modelRevision``

model_task: Optional = alias ``modelTask``: e.g. chat, completion or embedding.

inference_engine: Optional = alias ``inferenceEngine``: e.g. vllm, tgi, tei, ollama or llama_cpp.

api_format: Optional = alias ``apiFormat``

compute: Optional: The id of the LLMHostCompute that the LLMHost runs on.

gpu_count: Optional = alias ``gpuCount``

gpu_type: Optional = alias ``gpuType``

replicas: Optional

ready_replicas: Optional = alias ``readyReplicas``

status: Optional

status_message: Optional = alias ``statusMessage``

endpoint_url: Optional = alias ``endpointUrl``

public_url: Optional = alias ``publicUrl``

served_model_name: Optional = alias ``servedModelName``

is_active: Optional = alias ``isActive``

deployed_at: Optional = alias ``deployedAt``

cost_per_hour: Optional = alias ``costPerHour``

class smarter.models.llmhost.LLMHostStatus

Bases: smarter.models.base.SmarterCamelModel

An LLMHost’s state, read from Kubernetes now.

llmhost: str

status: str

message: Optional

replicas: Optional

ready_replicas: Optional = alias ``readyReplicas``

healthy: Optional

endpoint: Optional

public_url: Optional = alias ``publicUrl``

pods: list

checked_at: Optional = alias ``checkedAt``

class smarter.models.llmhost.LLMHostLogs

Bases: smarter.models.base.SmarterCamelModel

The latest log lines of an LLMHost’s pods.

llmhost: str

logs: JsonValue

class smarter.models.llmhost.LLMHostPlan

Bases: smarter.models.base.SmarterCamelModel

The Kubernetes resources that deploying an LLMHost would create, without creating them.

llmhost: str

resources: list

class smarter.models.llmhost.LLMHostReport

Bases: smarter.models.base.SmarterCamelModel

A report of every LLMHost that the user may read: counts, and estimated costs.

totals: dict

by_status: dict = alias ``byStatus``

by_engine: dict = alias ``byEngine``

llmhosts: list

class smarter.models.llmhost.DiscoveredModel

Bases: smarter.models.base.SmarterModel

A model found in a public catalog, e.g. Hugging Face, with the LLMHost manifest that would serve it.

catalog: Optional

source: Optional

repository: str

task: Optional

description: Optional

license: Optional

gated: Optional

downloads: Optional

likes: Optional

tags: list

manifest: Optional

class smarter.models.llmhost.DiscoveredManifest

Bases: smarter.models.base.SmarterModel

The response of discover/manifest/: a drafted LLMHost manifest.

manifest: dict

class smarter.models.llmhost.DiscoveredModels

Bases: smarter.models.base.SmarterModel

The models that a catalog search found.

models: list

`smarter.models.misc`

Models of secrets, prompt history, proxies, orchestrators and vectorsearches.

class smarter.models.misc.Secret

Bases: smarter.models.base.SmarterResource

A secret’s metadata.

The api never returns a secret’s value.

expires_at: Optional = alias ``expiresAt``

last_accessed: Optional = alias ``lastAccessed``

class smarter.models.misc.Orchestrator

Bases: smarter.models.base.SmarterResource

An orchestrator: a multi-agent workflow that composes LLMClients.

class smarter.models.misc.Vectorsearch

Bases: smarter.models.base.SmarterResource

A vectorsearch: a retrieval pipeline over one or more vectorstores.

class smarter.models.misc.PromptSession

Bases: smarter.models.base.SmarterModel

A chat session: one conversation with an LLMClient, by its session key.

id: int

created_at: Optional

updated_at: Optional

name: Optional

description: Optional

session_key: Optional

ip_address: Optional

user_agent: Optional

url: Optional

user_profile: Optional

llmclient: Optional

class smarter.models.misc.PromptToolCall

Bases: smarter.models.base.SmarterModel

A tool call that the LLM made in a chat session.

id: int

prompt: Union

created_at: Optional

updated_at: Optional

plugin: Optional

function_name: Optional

function_args: Optional

request: JsonValue

response: JsonValue

class smarter.models.misc.PromptPluginUsage

Bases: smarter.models.base.SmarterModel

A use of a plugin in a chat session.

id: int

prompt: Union

created_at: Optional

updated_at: Optional

plugin: Optional

input_text: Optional

class smarter.models.misc.Proxy

Bases: smarter.models.base.SmarterCamelModel

A proxy: authenticated passthrough access to an LLM provider’s own api.

The platform holds the provider’s api key. Callers use their Smarter api key instead, and may reach only the proxy’s allowed_paths.

name: str

description: Optional

provider: Optional

url: Optional: The proxy’s base url on the Smarter platform.

upstream_url: Optional = alias ``upstreamUrl``

allowed_paths: list = alias ``allowedPaths``

is_active: Optional = alias ``isActive``

class smarter.models.misc.Proxies

Bases: smarter.models.base.SmarterModel

The proxies that the user may call.

proxies: list

class smarter.models.misc.Provider

Bases: smarter.models.base.SmarterCamelModel

An LLM provider that the platform is configured for, e.g. openai.

name: str

description: Optional

base_url: Optional = alias ``baseUrl``

is_active: Optional = alias ``isActive``

class smarter.models.misc.ProviderModel

Bases: smarter.models.base.SmarterCamelModel

A model of an LLM provider.

name: str

description: Optional

max_tokens: Optional = alias ``maxTokens``

supports_streaming: Optional = alias ``supportsStreaming``

supports_tools: Optional = alias ``supportsTools``

metadata: Optional

Models

smarter.models.base

smarter.models.manifest

smarter.models.platform

smarter.models.account

smarter.models.llmclient

smarter.models.plugin

smarter.models.guardrail

smarter.models.mcpclient

smarter.models.vectorstore

smarter.models.llmhost

smarter.models.misc

`smarter.models.base`

`smarter.models.manifest`

`smarter.models.platform`

`smarter.models.account`

`smarter.models.llmclient`

`smarter.models.plugin`

`smarter.models.guardrail`

`smarter.models.mcpclient`

`smarter.models.vectorstore`

`smarter.models.llmhost`

`smarter.models.misc`