Skip to content

SequrityClient API Reference

This section provides the API reference for SequrityClient, the Python interface for Sequrity's Control API.

Overview

The Control API enables secure LLM interactions with policy enforcement. Key capabilities:

  • Chat Completions: OpenAI-compatible chat API with security features (toxicity filtering, PII redaction, topic guardrails)
  • Responses API: OpenAI Responses API with function calling, multi-turn, reasoning, and streaming support
  • Anthropic Messages: Anthropic Messages API with security features
  • LangGraph Integration: Execute LangGraph workflows with security policies via Sequrity's Dual-LLM runtime
  • Policy Generation: Generate SQRT policies from natural language descriptions

API Modules

Module Description
Chat Completion Chat completion API, request/response types, and result schemas
Responses API OpenAI Responses API, request/response types, and streaming events
Anthropic Messages Anthropic Messages API, request/response types
LangGraph sequrity mode LangGraph execution API and related types
Policy Generation Generate SQRT policies from natural language
ValueWithMeta Value wrapper with metadata for policy enforcement
SQRT SQRT policy parser and utilities
Integrations - LangGraph LangChain/LangGraph framework integration
Integrations - OAI Agents SDK OpenAI Agents SDK framework integration

Headers

Configuration headers for customizing security behavior:

SequrityClient 

SequrityClient(
    api_key: str,
    *,
    base_url: str | None = None,
    timeout: int = 300,
    control: ControlConfig | None = None,
)

Synchronous client for the Sequrity API.

Thin orchestrator that owns the HTTP connection pool and delegates to product-specific namespaces.

Example 
from sequrity import SequrityClient
from sequrity.control import ControlConfig, FeaturesHeader, SecurityPolicyHeader

client = SequrityClient(
    api_key="sq-xxx",
    control=ControlConfig(
        llm_api_key="sk-xxx",
        provider="openrouter",
        features=FeaturesHeader.dual_llm(),
        security_policy=SecurityPolicyHeader.dual_llm(),
    ),
)

response = client.control.chat.create(
    messages=[{"role": "user", "content": "Hello!"}],
    model="gpt-5-mini",
)

Parameters:

  • api_key

    (str) –

    Your Sequrity API key for authentication.

  • base_url

    (str | None, default: None ) –

    Sequrity API base URL. Defaults to the SEQURITY_BASE_URL environment variable, or https://api.sequrity.ai.

  • timeout

    (int, default: 300 ) –

    Default request timeout in seconds. Defaults to 300.

  • control

    (ControlConfig | None, default: None ) –

    Configuration for the Sequrity Control product. When omitted, an empty ControlConfig is used (all defaults are None, configure per-request instead).

Attributes:

  • control –

    Sequrity Control product namespace.

control instance-attribute 

control = ControlClient(_http_client, _api_key, _base_url, config=control)

Sequrity Control product namespace.