Analytics for
AI Agents
Track every run, step, model call, and dollar your agents spend.
One line of code. Zero config. Instant dashboard.
Works with LangChain, OpenAI Agents SDK, CrewAI, and any custom agent
Total Runs
12,847
+18% vs last 30d
Success Rate
94.2%
+2.1% vs last 30d
Avg Latency
1.2s
-340ms vs last 30d
Total Cost
$127.40
-12% vs last 30d
Runs over time
Recent Runs
Works with every framework
Everything you need to understand your agents
Lightweight, open-source observability built specifically for AI agent workflows.
Run Analytics
See success rates, latency, cost breakdowns, and trends across all agents and environments.
Step-level Traces
Every model call, tool invocation, and decision logged with inputs, outputs, tokens, and cost.
Cost Tracking
Know exactly what each agent run costs. Break down by model, tool, agent, or time period.
Error Monitoring
Catch failures fast. See top errors, failure rates, and drill into the exact step that broke.
Shareable Traces
Share any run trace with a public link. Great for debugging, reviews, and team collaboration.
Environment Aware
Separate metrics for dev, staging, and production. Compare behavior across environments.
API Key Auth
Authenticate agents via API keys. Dashboard uses session cookies. Zero friction either way.
Model Comparison
Compare model performance, cost, and reliability side-by-side across your agent fleet.
One Line Setup
Initialize with your API key. Call trace() or startRun(). That's it. Data flows instantly.
Three steps. Five minutes.
No complex config files. No YAML. No infra to manage.
Install the SDK
$ npm install @aspectbase/sdkAdd one line
import { AspectBase } from
"@aspectbase/sdk";
const ab = new AspectBase({
apiKey: "ab_live_..."
});Trace your agents
const run = await ab.startRun({
agentId: "research-bot",
environment: "prod"
});
// ... your agent logic
await run.end({ status: "success" });Why AspectBase?
Built for the agent era. Not retrofitted from the request era.
Simple, honest pricing
Free for small teams. Predictable as you scale.
Free
For side projects and experimentation
- ✓10,000 runs/month
- ✓7-day data retention
- ✓3 agents
- ✓Community support
Pro
For teams shipping agents to production
- ✓100,000 runs/month
- ✓90-day data retention
- ✓Unlimited agents
- ✓Priority support
- ✓Team members
- ✓Alerting & webhooks
Enterprise
For large teams with compliance needs
- ✓Unlimited runs
- ✓Custom retention
- ✓SSO / SAML
- ✓Dedicated support
- ✓SLA guarantee
- ✓Self-hosted option
Stop guessing. Start observing.
Your agents are making decisions, calling models, spending money. You should know what they're doing.