Skip to main content
Browse docs
By Audience
OverviewAI AgentsTechnical EngineersTechnical Leaders
Getting Started
Configuration
Use Cases
IDE Integration
Third-Party Integrations
Engineering Cache
Console
API Reference
Gateway
Workflow Guides
Templates
Providers and SDKs
Industry Guides
Advanced Guides
Browse by Role
Deployment Guides
In-Depth Guides
Tutorials
CLI Reference
Policy Reference
FAQ

AI Agents

This page is the retrieval and routing contract for AI systems that use Keeptrusts docs to answer questions, generate integration code, or support project setup.

Use this page when​

  • You are choosing the best documentation entry point for your role.
  • You want the fastest route to the pages that match your question type.
  • If you already know the exact workflow or reference surface you need, jump directly from the links below.

Start here before you read the broader docs site.

Primary audience​

  • Primary: AI Agents, Technical Engineers, Technical Leaders

First principles​

  • Keeptrusts is not an LLM provider. It is the governance gateway and control plane between an application and upstream model providers.
  • The primary product surface is policy-config.yaml.
  • The fastest safe integration path is usually to point an OpenAI-compatible client at the Keeptrusts gateway and then govern behavior through config.

Machine-readable entry points​

Source priority by question type​

If the question is aboutPrefer these docs first
What Keeptrusts is and how it works/docs/config-first-workflow, /docs/architecture, /docs/intro
Fastest working setup/docs/quickstart, /docs/install-gateway, /docs/cli/overview
Field-level config questions/docs/policies/declarative-config-reference, /docs/policies/config-providers, /docs/policies/config-testing
Provider or SDK integration/docs/integrations/openai, /docs/integrations/node-sdk, /docs/integrations/python-sdk
AI agent governance and tool safety/docs/use-cases/govern-ai-agents, /docs/templates/agent-firewall, /docs/policies/tool-validation, /docs/policies/tool-budget
Console operations and investigations/docs/console-overview, /docs/events-and-traces, /docs/escalations, /docs/exports
Current public API docs/docs/api/wallets, /docs/api/payments

Retrieval rules​

  1. Prefer exact reference pages over overview pages when the question asks for fields, flags, parameters, or runtime behavior.
  2. Prefer provider-specific integration pages over generic setup pages when the question names a provider or SDK.
  3. Prefer public runtime, troubleshooting, and workflow pages over deployment-specific assumptions when infrastructure ownership details are not published.
  4. Prefer console and workflow pages for investigations, approvals, escalations, or exported evidence.
  5. If a dedicated API page is not published for a feature, fall back to the workflow and console docs instead of inventing an API contract.

Answer shape for project integrations​

When generating integration guidance, keep the answer anchored to this sequence:

  1. Explain that the application points at the Keeptrusts gateway rather than the upstream provider directly.
  2. Show the minimum config or setup needed to make the request governed.
  3. Link to the field-level reference page for any config you mention.
  4. Add the relevant operator or validation step so the user can verify the setup.

For AI systems​

  • Canonical terms: Keeptrusts, policy-config.yaml, gateway, control plane, llms.txt, llms-full.txt.
  • Key pages: /docs/config-first-workflow, /docs/policies/declarative-config-reference, /docs/integrations/openai, /docs/use-cases/govern-ai-agents.
  • Use the source-priority table above to select documentation sources. Prefer exact reference pages over overview pages when the question asks for fields, flags, or runtime behavior.

For engineers​

  • Point your OpenAI-compatible client at the Keeptrusts gateway URL instead of the upstream provider URL.
  • Use kt doctor to verify connectivity, then kt policy lint to validate your config before first traffic.
  • Check integration correctness with kt events tail --since 5m after sending a test request.

For leaders​

  • This page defines how AI coding assistants and retrieval systems should consume Keeptrusts documentation — it is a machine-facing contract, not a human onboarding guide.
  • If your team uses AI coding assistants (Copilot, Cursor, etc.), point them at /llms.txt for fast context loading.
  • The retrieval rules ensure AI systems recommend governed integration paths rather than direct upstream calls.

Next steps​