One total
AI spend next to cloud spend.
Direct API bills and the AI line items already inside your GCP, AWS, and Azure invoices, in one view with one monthly AI budget.
All modules
AI spend
AIPulsenew
- Works with
- OpenAI
- Claude
- Agent Platform
- Bedrock
- Azure OpenAI
Every model, every token, every dollar
AIPulse brings OpenAI, Claude, Gemini Enterprise Agent Platform, Bedrock, and Azure OpenAI spend into one view. Cost and tokens per model and API key, live usage by the minute, and efficiency grades that show where to save.
- Every AI sourceOpenAI, Claude, Claude Enterprise, Agent Platform, Bedrock, Azure OpenAI.
- Down to the keyCost and tokens per model, project, workspace, and API key.
- Live token usageOne-minute buckets, split by uncached, cached, cache-write, and output.
- Efficiency gradesPrompt caching, batch usage, and model mix, with the saving attached.
Walkthrough
Inside AIPulse.
The components that matter: spend by source, top models, live token stream, efficiency grades, and what changed this week.
Spend by Source
This month · direct APIs and cloud bills
$2,042
↗ +11% vs last month
Claude EnterpriseAnalytics key$1284
Claude APIAdmin key$413
OpenAI APIAdmin key$178
Agent PlatformGCP billing$72
Azure OpenAIAzure billing$54
BedrockAWS billing$41
Top Models
Ranked by spend · blended cost per 1M tokens
claude-opus-5Claude Enterprise$15.20$812
claude-sonnet-5Claude Enterprise$3.71$472
claude-opus-5Claude API$2.83$198
claude-sonnet-5Claude API$0.88$186
gpt-5.6-terraOpenAI API$1.75$118
Token Stream
One-minute buckets · last hour
84.2k tokens/minoutput 3.1x baseline
Efficiency
Only levers with data are graded
B
Prompt caching
61% of input tokens served from cache
C
Batch usage
9% of API tokens at batch pricing
B
Model mix
48% of spend on frontier models
What Changed
Last 7 days
Output tokens spiked to 3.1x baseline
agents-prod · claude-opus-5 · Aug 28, 14:07
3 new seats activated
Claude Enterprise · Aug 27
New API key started billing
acme · portal-api-v2 · Aug 27
Agent Platform spend up 52% week over week
core-platform · claude-sonnet-5 · Aug 26
The full experience
The full AIPulse tour.
What you see below is a preview. Share your details and we'll email you a private link to the full interactive tour of AIPulse: every module, click-through, at your own pace.
Interactive product tour
Take the AIPulse tour yourself
An interactive, click-through tour of every AIPulse module. Explore at your own pace. Drop your details and we'll email you a private link. No call required. (Prefer a live walkthrough with an engineer? Book a guided demo.)
- No call required
- Explore at your own pace
- All the modules
Why this exists
“The AI bill doesn't arrive in one place. It lands in three consoles and two cloud invoices, and nobody owns the total.”
Every platform lead shipping LLM features
Live
See the spike while it's happening.
Token usage streams in by the minute, so a runaway agent shows up as a 3.1x output spike in minutes, not as a line item on tomorrow's cost report.
Actionable
Know which lever to pull.
Cost per 1M tokens by model and source, a Frontier / Mid / Light model mix, and a What Changed feed for new keys, new models, and new seats.
Sources
Every place your teams spend on AI.
Connect the AI organizations your engineers bill against. Managed AI on your cloud bills comes along automatically.
OpenAI API
Metered API spend across every project in your OpenAI organization.
- Cost by project, model, API key, and line item
- Tokens by project, API key, user, and model
- Cached tokens, batch flag, and service tier
- Existing spend limits imported as budgets
Claude API
Metered API spend across every workspace in your Claude Platform organization.
- Cost by workspace, model, and line item
- Tokens by workspace, API key, model, and service tier
- Cached, cache-write, and output tokens
- Claude Code cost per user
Claude Enterprise
Per-seat cost and usage across chat, projects, Claude Code, and Cowork.
- Cost per user, product, and model
- Tokens, including cached and cache-write
- Request counts and engagement
- Inactive paid seats flagged
Already on your cloud bill
No setup, no extra keys.
- Gemini Enterprise Agent PlatformFormerly Vertex AI · from your GCP billing export
- Amazon BedrockFrom your AWS Cost and Usage Report
- Azure OpenAIFrom your Azure Cost Management export
Trust
Counts and costs. Never content.
Reads usage and cost only
Nothing is written back, and no model calls are made on your behalf.
Keys stored encrypted
Provider keys are encrypted at rest. OpenAI read-only admin keys are supported.
Never reads conversations
AIPulse sees counts and costs. Prompt and response content is never collected.
Per-user cost stays private
Named, per-person spend is visible to Pulse Admins only. Everyone else sees totals.
Same read-only posture as the rest of Pulse.
Questions
AIPulse, answered.
Live walkthrough
See Pulse in action,
today!
We walk you through every Pulse module (BillPulse, AIPulse, SecureMonitor, PulseArc, Ollie, Killswitch) on a representative cloud and AI environment loaded with realistic data, together with one of our cloud experts. About 30 minutes.
What you'll see in the demo
- 01Daily cost intelligence for GCP, AWS, and Azure, with OpenAI and Claude API spend by model and workspace alongside.
- 02Severity-ranked GCP security findings with the matching AI remediation plan one click away.
- 03An interactive PulseArc map of resources with security overlay: a live document, not a drawing.
- 04Ollie answering plain-English questions about a realistic multi-cloud environment, with sources.
hello@zenta.ai~24h reply
Free Cloud Security AssessmentSeparate offering
A one-off, no-charge audit that surfaces critical SCC findings in your GCP, a complimentary service to help teams understand their security gaps before adopting Pulse.
Request your free assessment