Solutions / Observe
IRIS8 Observe

See every AI interaction across your company — in minutes.

Point your apps, agents, and AI tools at the IRIS8 AI Gateway — one OpenAI-compatible endpoint — and your entire AI estate becomes visible: every model, every tool, every prompt, every team, every dollar.

Available nowSelf-servePowered by the AI Gateway
app.iris8.ai/observe
143AI systems
27shadow
$48.2Kspend / mo
commercial-underwriting-v3frontier-lm-7 · lendingActive
support-copilotgpt-class · contact centerActive
unregistered: cursor-agentengineering · 14 seatsShadow
5 min

median time from signup to first visibility

100%

coverage of traffic through the gateway

20+

providers, frameworks, and MCP integrations

0

agents to install, surveys to chase

The problem

Every company has an AI estate. Almost none can see it.

Unknown estate

Teams adopt models, agents, and AI tools faster than any registry can track — 60% of enterprise AI is undocumented.

Untracked spend

AI costs scatter across provider invoices, team cards, and SaaS bundles. Nobody owns the number.

Survey-based inventories

Annual questionnaires produce stale spreadsheets. Your estate changes daily; your inventory doesn’t.

Capabilities in depth

From first request to full picture.

The gateway

One gateway, total coverage

Swap a base URL and keep your code. The gateway speaks the APIs your teams already use — and everything that flows through it becomes governable.

  • OpenAI-compatible endpoint; drop-in for existing clients
  • Plugins for LiteLLM, Portkey, Kong, and Azure APIM
  • SDK hooks for LangGraph, CrewAI, AutoGen, and MCP
  • Sub-100ms passthrough — invisible to users
quickstart — 5 minutes
// point your existing client at the gateway
client = OpenAI(
  base_url="https://gw.iris8.ai/v1",
  api_key=env("IRIS8_KEY")
)
// done. every call is now visible, secured, and provable.
Inventory

A living inventory, not a stale spreadsheet

Every model, agent, tool, and MCP connection registers itself from real traffic — with lifecycle states, ownership, and autonomy tiers attached automatically.

  • Auto-registration on first request; shadow systems flagged
  • Lifecycle states from first-seen to retired
  • Ownership and team attribution from identity
  • Certification states and autonomy tiers, governance-ready
app.iris8.ai/observe/inventory
claims-triage-agentcrewai · claims opsActive
kyb-research-toolmcp · onboardingNew
marketing-draftersaas ai · growthShadow
fraud-screening-v2self-hosted · riskActive
sql-copilotide plugin · data engShadow
Analytics

Numbers your CFO and CISO both want

Live usage, latency, and spend by team, project, model, and provider — with the drill-downs to answer "who, what, and how much" in one click.

  • Spend and token analytics with provider mix
  • Latency and error tracking per route
  • Scheduled reports and export anywhere
  • Anomaly flags when usage patterns shift
app.iris8.ai/observe
143AI systems
27shadow
$48.2Kspend / mo
commercial-underwriting-v3frontier-lm-7 · lendingActive
support-copilotgpt-class · contact centerActive
unregistered: cursor-agentengineering · 14 seatsShadow
Everything included

The full capability set.

You can't govern AI you can't see. Now you can see all of it.

Gateway

One drop-in endpoint

Swap your base URL, keep your code. The gateway speaks the APIs your teams already use, with SDK hooks for LangGraph, CrewAI, AutoGen, and MCP — or runs as a plugin behind LiteLLM, Portkey, Kong, and Azure APIM.

Inventory

A live estate map

Models, agents, tools, and MCP connections inventoried automatically from real traffic — not from a survey nobody answers. Lifecycle states from first-seen to retired.

Discovery

Shadow-AI discovery

The tools and models in use that nobody registered — surfaced by team, by app, and by data sensitivity, the moment they touch the gateway.

Analytics

Usage & spend analytics

Requests, tokens, latency, and cost by team, project, model, and provider — live dashboards, scheduled reports, export anywhere.

Graph

The dependency graph

Which agents call which tools, touch which data, run on which models — blast radius answered before anyone asks.

Ready

Governance-ready from day one

Every discovered system carries a certification state and an autonomy tier — so when you're ready for enforcement with Gate, the foundation is already in place.

How it works

Visibility in three steps.

  1. Point your traffic at the gateway. Five-minute setup, free tier, no procurement meeting required.
  2. Watch the estate light up. Inventory, usage, spend, and shadow AI appear as traffic flows — no agents to install, no surveys to chase.
  3. Invite risk and compliance. Hand them a live estate map instead of a stale spreadsheet. They'll thank you.
Questions

The short answers.

Do we have to replace our existing gateway?

No. IRIS8 runs as a drop-in endpoint or as a plugin behind LiteLLM, Portkey, Kong, or Azure APIM — whichever path is least disruptive.

Does visibility slow down requests?

Observation is passive on the request path — sub-100ms passthrough at p95. Your users won’t notice; your dashboards will.

What about AI used outside the gateway?

Gateway traffic is the wedge. Repo and cloud scanning on the Control tier extends discovery to systems that never touch the gateway.

Your whole AI estate, visible by Friday.

Free to start. No sales call required.

Start free