contextwall-sdk 0.1.4

ContextWall SDK - drop-in context firewall enforcement for any LLM

contextwall-sdk

Drop-in context firewall enforcement for any LLM. Routes calls through a local ContextWall daemon, enforcing your policy before content reaches the model.

Built-in wrappers for Anthropic and OpenAI. Any OpenAI-compatible API (Mistral, Groq, Together, Ollama, etc.) works via SafeOpenAI. For other providers, use the zero-code proxy mode.

Install

pip install 'contextwall-sdk[anthropic]'   # Anthropic
pip install 'contextwall-sdk[openai]'      # OpenAI + any OpenAI-compatible API
pip install 'contextwall-sdk[all]'         # both

Usage

Anthropic:

from contextwall_sdk import SafeAnthropic, ContextWallBlockedError

client = SafeAnthropic(api_key="sk-ant-...", ctxfw_url="http://localhost:8080")
# use exactly like the standard Anthropic client

OpenAI / any OpenAI-compatible API:

from contextwall_sdk import SafeOpenAI

# OpenAI
client = SafeOpenAI(api_key="sk-...", ctxfw_url="http://localhost:8080")

# Mistral, Groq, Together, Ollama — just pass base_url
client = SafeOpenAI(api_key="...", ctxfw_url="http://localhost:8080",
                    base_url="https://api.mistral.ai/v1")

Zero-code proxy mode (any provider, no SDK changes):

export ANTHROPIC_BASE_URL=http://localhost:8080/proxy/anthropic
export OPENAI_BASE_URL=http://localhost:8080/proxy/openai
# your existing code runs unchanged

Blocked requests raise ContextWallBlockedError with the policy violation detail.

Daemon setup

Requires a running ContextWall daemon. See the quickstart or the GitHub repo.