Platform overview
Connect and build
Explore how DT uses internal APIs and MCP tools, and how your applications can work with DT’s persistent agents.
Explore integrations and APIsDecision Terrain Resource Center
AI platform guides.
Tools for your next decision.
Explore how to connect, evaluate, and operate AI systems. Find DT platform capabilities, browser-only planning tools, and technical guides, including detailed guidance for defense, intelligence, and air-gapped environments.
DT platform capabilities
Understand how DT fits your systems.
Start with the product details, then use the guides and planning tools below to frame the work, assess the evidence, and prepare your implementation.
Platform overview
Choose a deployment
Compare running DT on your hardware, in your cloud account, or air-gapped. Establish where models, data, and execution operate.
Review deployment optionsPlatform overview
Evaluate the controls
Examine identity, tool access, action-specific approvals, and available audit evidence for the proposed configuration.
Review security and governanceBrowser-only planning tools
Structure the question before the commitment grows.
Each tool runs in your browser without submitting your answers. The five checklists save drafts on this device; the worksheets do not save entries. Keep all inputs unclassified and non-sensitive.
AI Facts · Public alpha
AI Facts: NIST AI RMF Checklist
Prepare for your next AI review. Document safeguards, evidence notes, and gaps with 40 core checks and six optional agent checks. Export an assessment report and AI Facts label.
Start an assessmentUnofficial NIST AI 600-1 checklist
Generative AI Profile Checklist
Review one generative AI use with 24 selected, unofficial checks mapped to NIST AI 600-1 suggested actions. Document evidence and gaps, then print or export a JSON backup.
Start a generative AI reviewUnofficial OWASP 2026 checklist
OWASP LLM Top 10 Checklist
Review one LLM application with 30 practical checks adapted from the OWASP Top 10 for LLM Applications 2026. Record evidence and gaps, then print or export a JSON backup.
Start an LLM security reviewUnofficial OWASP Agentic 2026 checklist
OWASP Agentic Top 10 Checklist
Review one agentic application with 30 practical checks adapted from the OWASP Top 10 for Agentic Applications 2026. Record evidence and gaps, then print or export a JSON backup.
Start an agentic security reviewUnofficial secure AI development checklist
NIST SP 800-218A Checklist
Review secure AI model development with 34 selected, unofficial checks mapped to NIST SP 800-218A tasks across Prepare, Protect, Produce, and Respond.
Start a secure development reviewInteractive worksheet
Mission AI Readiness Worksheet
Assess six conditions that shape whether an AI opportunity is ready for focused testing, from mission ownership and data to governance and adoption.
Open the worksheetInteractive worksheet
AI Technology Evaluation Scorecard
Compare two technologies against weighted mission, technical, security, integration, lifecycle, and evidence criteria without sending data anywhere.
Open the worksheetInteractive worksheet
AI Prototype Test-Plan Builder
Turn a central uncertainty into a bounded, printable test plan with measures, evidence needs, failure conditions, and a decision gate.
Open the worksheetTechnical and decision guides
Plan integration, evaluation, and operations.
Explore architecture, model updates, implementation handoff, readiness, and evaluation. The defense-focused guides draw on public CDAO, NIST, GAO, CISA, and defense sources and retain their specific mission context.
Governance guide
AI Agent Governance: Identity, Tools, Actions, and Evidence
Control delegated authority with a practical framework for identity, tool scope, runtime policy, human oversight, and reviewable evidence.
Read the guideReadiness guide
What AI Readiness Means for a Defense Program
A practical definition of readiness that connects mission value, technical conditions, governance, people, and evidence.
Read the guideEvaluation guide
How to Evaluate an AI Vendor Beyond the Demonstration
A structured method for testing claims against the workflow, operating environment, lifecycle, and evidence the decision actually requires.
Read the guidePrototyping guide
Prototype vs. Pilot: What Evidence Do You Need First?
Use the uncertainty, user involvement, environment, and next decision to choose the right learning vehicle before scope expands.
Read the guideArchitecture guide
Air-Gapped Agentic AI Architecture: A Practical Reference
Map the inference, orchestration, identity, tools, data, observability, evaluation, and update layers needed inside a disconnected boundary.
Read the guideOperations guide
Model Updates in Air-Gapped AI: A Controlled Lifecycle
Plan controlled intake, provenance, evaluation, approval, deployment, monitoring, rollback, and recovery for models and their dependencies.
Read the guidePlatform evaluation guide
How to Evaluate an Open-Source Agent Platform for an Air Gap
Assess offline operation, inference portability, permissions, auditability, supply-chain evidence, sustainment, and exit options.
Read the guideDemo
Insider Risk Alert Triage With an AI Agent
Watch an agent gather the evidence for an after-hours alert, wait for an analyst to approve the case, and get blocked from tipping off a manager.
Read the guideVideo
Decision Terrain Platform Overview
See agents work in chat, through the API, and in your systems, then the rules, approvals, halts, and records that let your CISO say yes.
Read the guideDemo
AI Agent Blast Radius
See what one agent could reach if it were compromised, how each path is graded, and how removing one tool shrinks its reach.
Read the guideDemo
AI Agent Halt and Release
Watch a Splunk alert halt one agent through a scoped API, the record show what was enforced, and a named person release it.
Read the guideHow to use the library
Frame. Compare. Test. Decide.
Start with the uncertainty that matters. Use the matching tool to organize it and a guide to examine the assumptions. Your team can lead implementation, with optional DT engineering and supportavailable where you need additional expertise.
Frame
State the mission decision, user, constraints, and consequence of being wrong.
Compare
Make criteria and tradeoffs explicit before a product or approach becomes the default.
Test
Design the smallest credible experiment that can answer the central question.
Decide
Use the evidence to advance, revise, gather more information, or stop.