Lightrun AI🔗
Lightrun AI enables AI agents to investigate application behavior using Lightrun's production-safe runtime data. It brings retained and live runtime context into AI-driven debugging workflows, without requiring code changes, rebuilds, or redeployments.
Lightrun AI is built around two layers:
- Lightrun MCP connects AI agents to Lightrun runtime capabilities through the Model Context Protocol.
- Lightrun AI Skills provide reusable investigation workflows that guide AI agents through focused, evidence-first debugging with Lightrun MCP.
The typical flow is: a developer asks an AI agent to investigate an issue, the agent follows a Lightrun AI Skill, the skill guides the agent through Lightrun MCP, and MCP searches retained Runtime History or collects fresh runtime evidence from the target application.
Lightrun MCP🔗
Lightrun MCP lets AI agents such as Cursor, GitHub Copilot, Claude Code, Gemini, Devin, and other MCP-compatible tools use Lightrun's runtime context on your behalf.
With Lightrun MCP, an AI agent can search retained snapshot hits through Runtime History or collect fresh expression values, call stacks, execution duration, execution counts, and custom metrics from live applications. This allows the agent to validate hypotheses against real application behavior instead of relying only on static code analysis.
To get started, see:
- What is the Lightrun MCP?
- MCP quickstart guide
- Use Runtime History with Lightrun MCP
- Supported Lightrun MCP tools
Lightrun AI Skills🔗
Lightrun AI Skills are reusable instructions for AI agents. They define repeatable workflows for using Lightrun MCP during runtime investigations, including preflight checks, evidence collection, diagnosis, and handoff output.
Skills help AI agents use Lightrun consistently. Instead of asking an agent to "debug this" and relying on a generic workflow, you can use a Lightrun skill to guide the agent through a structured investigation that starts with runtime source discovery and ends with a concrete diagnosis and next action.
Available skills include the Live Runtime Debugging Skill, the Error Remediation Automation Skill, the Ask Prod Skill, the Slow Execution Diagnosis Skill, and the Runtime-Aware PR Review Skill.
To learn more, see Lightrun AI Skills.
AI SRE🔗
For Lightrun's AI SRE documentation, see AI SRE.