AgentPulse gives you instant alert rules, cost-per-product breakdowns, and execution traces for every AI agent — all ingested via a single HTTP call with no SDK required.
Fully hosted & managed · No SDK required · Live in minutes
From per-product spend breakdowns to instant alert rules — a full observability stack with no SDK or instrumentation changes.
Bar chart of USD spend per product with % share of total. Click any product to filter the entire dashboard — know exactly what's burning budget before your bill arrives.
5 configurable metrics: error_rate, success_rate, avg_latency_ms, failed_runs, cost_usd. Set gt/lt thresholds with rolling windows — fires to Slack, webhook, or email.
POST /v1/runs with a Bearer secret. No SDK, no library, no agent code changes. Includes status, duration, steps, tokens, and cost.
Full org isolation — each organisation sees only its own agents, runs, and alerts. Invite team members, assign admin/member roles, and manage products per org.
Step-by-step breakdown of every run — LLM calls, tool calls, retrievals, and reasoning steps with per-step timing, token usage, and error context.
Real-time run volume charts, agent fleet health, success rate, avg latency, and active agent count. 24h / 7d / 30d period selector across all views.
Create your org at agentpulse.sentinelflux.in — fully hosted and managed. Nothing to deploy or maintain.
Register an agent in the dashboard to get a Bearer secret, then POST one curl after each agent run. No SDK, no library.
Set alert rules on error rate, cost, or latency. Watch cost-by-product on the dashboard. Get notified on Slack or webhook the moment a threshold is breached.
Fully hosted and managed by us — nothing to deploy. No credit card required for the Free tier.
Get full visibility into every run in under 10 minutes.