Learning center
Understand the control before you connect a model.
Start with a captioned walkthrough under three minutes, read the plain-language basics, or go directly to an operator workflow. Every resource states what works today and what remains your responsibility.
- Plain-language starting path
- Captioned product walkthroughs
- Operator SOPs with current boundaries
Choose your path
Start at your level. Finish with a working boundary.
Each path moves from an understandable outcome to the exact setup, validation, and responsibility split needed to use it safely.
Topic
20 resources
Start here
AI Gateway HQ: the control boundary in under three minutes
Watch the working product turn provider access, routes, failover, policy, hard budgets, managed billing, and evidence into one understandable operating path.
Start here
Send your first governed AI request
Follow the shortest safe path from approved model supply to a bounded route, hard budget, one-time workload identity, test request, and reason-coded evidence.
Start here
Connect your provider API key safely
Keep one encrypted provider credential at the gateway boundary, verify it without inference, and give applications separate workload identities and approved routes.
Start here
Can an AI gateway use your ChatGPT or Claude subscription?
Understand what BYOK means, when a consumer AI subscription can stay in use, and which credentials an AI gateway should never ask you to upload.
Start here
Start with prepaid managed Amazon Bedrock
See how purchased credit, a reviewed model catalog, explicit routes, hard budgets, and bounded reloads create a no-provider-key starting path.
Cost control
Set a hard AI budget before traffic starts
Set a workspace-wide ceiling, reserve exposure before provider dispatch, and keep managed credit reloads independently bounded.
Reliability
Configure provider failover without bypassing controls
Build fallback from independent approved supply, remove unhealthy targets automatically, and fail closed when nothing qualifies.
Cost control
Review and approve an AI cost insight
Move from retained metadata to one exact, version-bound routing change with human approval, audit evidence, rollback, and cautious outcome measurement.
Governance
Respond to an enterprise trust review safely
See how a bounded inquiry becomes a qualified, owner-approved evidence delivery without exposing sensitive contact or AI content.
Governance
AI gateway security features: a CISO buyer checklist
Compare identity, data protection, agent permissions, budget abuse, audit evidence, deployment isolation, and resilience before choosing an AI gateway.
Governance
Test an enterprise AI gateway before you buy it
Use 12 hands-on tests to learn whether an AI gateway will control access and spend, survive a provider failure, and preserve customer boundaries.
Start here
What Is an AI Gateway? A Plain-Language Guide
Learn how an AI gateway controls access, model routing, spending, and operating evidence between your applications and AI providers.
Start here
Connect Codex CLI to Groq with budgets and fallback
Route Codex CLI to Groq through AI Gateway HQ. Configure an isolated workload key, a hard budget, rate limits, a stable model alias, and tested fallback.
Reliability
Automatic AI provider failover without silent policy drift
See how health circuits, quota cooldowns, capacity evidence, and bounded fallback keep an outage from becoming an uncontrolled model change.
Cost control
AI gateway cost control: stop surprise spend before dispatch
A practical control plan for workload ownership, conservative reservations, provider limits, capped reloads, cost-aware routing, and invoice reconciliation.
Portfolio oversight
Private-equity AI value realization
Connect portfolio-company AI spend, adoption, business results, evidence quality, and operating risk without turning activity into a fictional return.
Portfolio oversight
A measurement system for AI across a private-equity portfolio
Compare AI spending, controls, adoption, and business results across portfolio companies without pooling their private data.
Governance
AI governance without storing every prompt
Learn what an AI gateway can record for policy, cost, and audit review without adding prompt or response text to its durable ledger.
Cost control
LLM cost attribution: reconcile provider bills
Compare provider financial totals with request-level gateway evidence, investigate variance, and allocate cost without uploading the source bill.
Start here
Use managed Bedrock with no provider API key
Fund a bounded wallet, activate the approved managed model, create a safe route, and stop automatically when purchased credit is gone.