Build with governed agent workflows.
Connect Contextus through the SDK, CLI, MCP, VS Code, or REST API.
⌘K
Contextus is the control plane for AI agent actions
Contextus compiles the right context before the model acts, governs risky actions, and keeps proof of what happened.
Option A - Python SDK (2 minutes)
1. Install
pip install contextus-sdk2. Set your API key
export CONTEXTUS_API_KEY="ctx-..."3. Compile your first context
from contextus import ContextusClient
client = ContextusClient() # reads CONTEXTUS_API_KEY from env
result = client.compile(
context="[paste your long context here]",
query="What are the key risks?"
)
print(result.compiled_context) # trimmed, relevant context
print(f"Reduced by {result.reduction_pct:.0f}%") # e.g. "Reduced by 71%"Option B - MCP (Cursor / Claude Desktop)
Install the Contextus MCP adapter and Cursor/Claude will automatically compile context before every LLM call.
Install via npm
npm install -g @contextus/mcp-adapterOption C - VS Code Extension
Install the Contextus extension from the VS Code Marketplace. Governed compile flows appear inline while you code.
Open in VS Code MarketplaceStill stuck?
Email us at info@coreledger.ca
Documentation Assistant
Ask about compile, policy, approvals, or rollout
Hi! I'm your Contextus documentation assistant. I can help you find information about setup, configuration, API usage, and more. What would you like to know?
08:01 AM
Documentation-only answers powered by Contextus RAG