ReskPoints

The AI Agent Logger

Track Every Action Your Agents Take

ReskPoints is a Python library that logs every action performed by AI agents -- function calls, tool executions, LLM completions -- with full context: probability, parameters, results, and timestamps. Ship logs to Datadog, Prometheus, OpenTelemetry, webhooks, or local files. Built for observability in production AI agent deployments.

Project Stats

500
PyPI Downloads
0
GitHub Stars
v0.1.0
Latest Version
Python
Language

Key Features

Multi-Destination Logging

Send agent logs to any observability backend without changing your code.

  • Datadog: Ship logs as Datadog events with custom attributes
  • Prometheus: Expose metrics for agent activity, latency, error rates
  • OpenTelemetry: Trace agent executions across distributed systems
  • Webhooks: HTTP callbacks for real-time agent monitoring
  • File: Local JSON or CSV logging for debugging

Rich Agent Context

Every log entry captures the full context of the agent action.

  • Function name and parameters
  • Token probabilities at decision points
  • Execution duration and timestamps
  • Success/failure status with error details
  • Session and trace identifiers for correlation

Minimal Performance Overhead

Designed for production use with asynchronous, non-blocking log shipping.

  • Async batch processing of log entries
  • Configurable buffer size and flush interval
  • Automatic retry on network failures
  • Zero-dependency logging core

Framework Agnostic

Works with any agent framework -- LangChain, CrewAI, AutoGen, or custom implementations.

  • Decorator-based instrumentation
  • Context manager for manual scoping
  • Middleware for framework-specific integrations
  • OpenTelemetry-compatible spans

Quick Install

pip install reskpoints

Usage Example

Basic Logging

from reskpoints import AgentLogger logger = AgentLogger( destination="datadog", api_key="your-datadog-key", ) @logger.track("send_email") def send_email(to: str, subject: str, body: str): # Your agent action here return {"status": "sent", "to": to} send_email(to="user@example.com", subject="Hello")

Multi-Destination

from reskpoints import AgentLogger logger = AgentLogger( destinations=["datadog", "prometheus", "file"], config={ "datadog": {"api_key": "...", "site": "datadoghq.eu"}, "prometheus": {"port": 8000, "namespace": "my_agent"}, "file": {"path": "./agent_logs.json", "format": "json"}, } )

Get Started with ReskPoints

Add observability to your AI agent deployments.

View on GitHub PyPI Package

Technical Support

For technical inquiries and integration support:

contact[@]resk.fr

Contact Our Team