Skip to main content

​
What is Agentflare?

Agentflare provides protocol-level observability for AI agent tool servers, giving you complete transparency and control over your AI systems. Track every decision, measure tool selection performance, and validate compliance in real-time.
  • Why did my agent make that decision?
  • What was the agent thinking at each step?
  • How much is each decision costing me?
  • Where are the performance bottlenecks?

​
Core Architecture

Agentflare operates on two foundational components:

​
1. Multi-Tenant Proxy

A hosted proxy service that sits between your AI agents and tool servers, providing:
  • Zero-configuration observability - Add your tool server and get a proxy URL
  • Custom domain support - Use your own domain (e.g., mcp.yourcompany.com)
  • Real-time tool reasoning capture - See why agents select specific tools
  • Automatic tracking - Captures reasoning, confidence, latency, and cost per decision

​
2. Agentflare Dashboard

A comprehensive web interface for visualizing agent behavior:
  • Live tool call feed - Real-time stream of all agent actions
  • Tool reasoning - See why agents selected specific tools
  • Agent reasoning - Step-by-step decision visualization
  • Schema Studio - Fine-tune decision pathways without redeploying code
  • Performance analytics - Identify bottlenecks and optimize
  • Cost breakdown - Detailed cost analysis per tool, session, and model

​
Key Features

​
Decision Transparency

Tool Reasoning

See exactly why your AI agents select specific tools

Agent Reasoning

Trace step-by-step agent decision-making processes

Memory Replay

Reconstruct complete agent decision sequences

Schema Studio

Fine-tune tool schemas and decision pathways in real-time

​
Performance Monitoring

  • Real-time Metrics
  • Historical Analysis
  • Alerting
  • Latency tracking - Response times for each tool call
  • Throughput monitoring - Requests per second, per minute
  • Error rates - Track failures and their causes
  • Resource usage - Memory, CPU, and network utilization

​
Cost Management

Track costs at multiple levels:
  • Per tool call - Individual operation costs
  • Per session - Complete conversation costs
  • Per model - Compare cost efficiency across models
  • Per user/workspace - Multi-tenant cost attribution
Set and monitor budgets:
  • Monthly/daily budgets - Control spending limits
  • Cost alerts - Get notified before exceeding budgets
  • Usage forecasting - Predict future costs
  • Cost optimization - Recommendations for reducing expenses
Analyze spending patterns:
  • Cost trends - Track spending over time
  • Tool efficiency - Identify most/least cost-effective tools
  • Model comparison - Compare cost vs. performance
  • ROI analysis - Measure return on AI investment

​
How It Works

​
Simple Proxy Integration

Getting started with Agentflare is straightforward:
1

Add Your Tool Server

Add your MCP tool server to Agentflare via the dashboard
2

Get Proxy URL

Receive a unique proxy URL for your tool server
3

Update Your Agent

Point your AI agent to use the Agentflare proxy URL
4

Start Observing

Instantly see tool calls, reasoning, and performance in real-time
# Example: Point your client to the Agentflare proxy
export MCP_SERVER_URL="https://proxy.agentflare.com/<workspace>/<server-slug>"

​
OTEL Export

Send data to your existing observability stack:
  • Jaeger
  • Datadog
  • New Relic
exporters:
  jaeger:
    endpoint: http://localhost:14268/api/traces
    headers:
      x-api-key: <your-jaeger-key>

​
Use Cases

​
Development & Debugging

  • Debug agent behavior - Understand why agents make specific choices
  • Optimize performance - Identify and fix bottlenecks
  • Test different prompts - A/B test prompt variations
  • Validate reasoning - Ensure agents think as expected

​
Production Monitoring

  • Real-time alerting - Get notified of issues immediately
  • Performance tracking - Monitor SLAs and response times
  • Cost control - Stay within budget and optimize spending
  • Compliance - Audit agent decisions for regulatory requirements

​
Business Intelligence

  • Usage analytics - Understand how agents are being used
  • ROI measurement - Calculate return on AI investment
  • Capacity planning - Plan for future growth
  • Competitive analysis - Compare different AI solutions

​
Getting Started

1

Sign up for Agentflare

Create your account at agentflare.com
2

Add your tool server

Navigate to Dashboard → Tool Servers → Add Server
3

Get your proxy URL

Copy the generated proxy URL for your tool server
4

Start monitoring

Point your agent to the proxy URL and see real-time observability

​
Next Steps

Quick Start

Get up and running in 5 minutes

Thought Tracing

Learn about decision visibility

Performance Analytics

Monitor and optimize performance

Cost Tracking

Control and optimize costs
Agentflare is designed to be minimally invasive with < 1ms latency overhead and asynchronous data collection.
⌘I