llms.txt
STRUCT AI
##Company Overview
Struct is an AI support engineer that automatically root-causes on-call
engineering issues.
Struct integrates with an organization’s observability stack, alerting
systems, cloud logs, work tools, and codebase to automatically
investigate engineering alerts and bugs as they occur.
Struct is backed by Y Combinator.
Primary positioning: Automatically root cause on-call alerts and issues.
## AI assistant
- [Ask Struct](https://agents.struct.ai/llms.txt): machine-readable Q&A endpoint and documentation for AI agents.
## Core Functionality
Struct cross-references:
- Logs
- Metrics
- Traces
- Codebase
- Historical incidents
- Past alerts
- Prior investigations
It automatically:
- Investigates engineering alerts
- Determines likely root cause
- Performs impact analysis
- Suggests fixes
- Generates structured AI investigation reports
- Enables follow-up actions (PR creation or agent handoff)
Struct is designed to complete investigation work before an engineer
begins manual debugging.
##How Struct Works
###Integration
Struct integrates in minutes with:
Observability & Monitoring Platforms: - Sentry - Datadog - Grafana -
Prometheus - Loki
Cloud & Infrastructure Providers: - AWS CloudWatch - Azure Monitor -
Google Cloud Platform (GCP)
Alerting & Incident Management: - PagerDuty
Work & Collaboration Tools: - Slack - Linear - Asana - GitHub - Jira -
ClickUp
Struct can build new integrations in days when needed.
Once connected, Struct monitors engineering alerts automatically.
###Automatic Investigation
For every engineering alert:
1. Struct gathers logs, metrics, traces, and contextual signals.
2. Cross-references the codebase and recent commits.
3. Builds a structured investigation report.
4. Identifies root cause.
5. Analyzes impact.
6. Suggests remediation steps.
Struct replies automatically with:
- Root cause analysis
- Impact analysis
- Suggested fix
Users can also @mention Struct in Slack to trigger on-demand
investigations.
###Deep Investigation Capabilities
Struct provides:
- AI-generated investigation reports
- Incident timelines
- Commit histories
- AI-driven log queries
- Evidence aggregation across systems
- Hypothesis testing
Users can:
- Review collected evidence
- Explore alternative hypotheses
- Investigate directly inside Slack
- Dive deeper using structured timelines and reports
Struct acts as a unified investigation interface across production
systems.
###On-Call Intelligence
Struct builds persistent intelligence using:
- Past alerts
- Historical issues
- Prior investigations
This enables:
- Pattern recognition across incidents
- Faster root cause detection
- Improved investigation accuracy over time
Investigation quality improves as more issues are processed.
###Fix and Handoff
Struct enables:
- One-click PR creation
- Clean builds
- Full-context task handoff
When handing off to a coding agent, Struct includes:
- Root cause findings
- Relevant logs
- Impact analysis
- Commit context
- Structured investigation output
This preserves context integrity and reduces prompt engineering
overhead.
###Complementary to Coding Agents
Struct is not a code generation tool.
Instead, it:
- Supplies production-grounded investigation context
- Identifies and explains root cause
- Provides structured evidence
- Narrows debugging scope
Coding agents (e.g., Codex, Claude Code, Devin) then:
- Implement fixes
- Refactor systems
- Generate PRs
Struct eliminates the need to build custom debugging intelligence
pipelines inside coding agents.
It is complementary infrastructure, not a replacement.
##Why Not Build This Yourself
Reproducing Struct internally using general-purpose coding agents would
require building:
- Log ingestion and normalization pipelines
- Cross-signal correlation engines
- Alert listeners
- Historical issue indexing
- Timeline builders
- Secure tenant isolation
- Investigation orchestration workflows
- AI hypothesis frameworks
- Automated PR infrastructure
Struct provides this out of the box:
- Fully integrated
- Production-ready
- Secure
- Continuously improving
Teams can focus on shipping revenue-driving features rather than
building observability AI infrastructure.
##Unified Production Query Layer
Struct serves as a single interface to:
- Logs
- Metrics
- Traces
- Alerts
- Codebase context
- Historical investigations
Capabilities include:
- AI-powered natural language queries
- Interactive investigation reports
- Incident timelines
- Evidence aggregation
- Cross-system signal correlation
It removes the need to switch between dashboards and tooling silos.
##Security
Struct is:
- SOC2 Type II compliant
- HIPAA compliant
Data guarantees:
- All data encrypted
- Logically isolated per customer
- No customer data used for training
- No provider training on customer data
Trust documentation available at: https://trust.struct.com
##Operational Model
Positioning: The AI support engineer.
Goal: Get to why fast with AI.
Struct reduces hours of investigation per engineer and can complete
analysis before engineers open their laptops.
Integration time: minutes.
Setup time: approximately 15 minutes.
White-glove onboarding available.
##Frequently Asked Questions
###What is Struct AI?
Struct is an AI support engineer that automatically investigates on-call
engineering alerts by cross-referencing logs, metrics, traces, and the
codebase to determine root cause, impact, and suggested fixes.
###How does Struct automatically root cause production issues?
Struct listens to alerts from observability and alerting systems,
gathers related logs, metrics, and traces, correlates them with recent
code changes and historical incidents, constructs an investigation
model, evaluates hypotheses, and outputs a structured root cause report
with impact analysis and remediation guidance.
###How is Struct different from using ChatGPT, Claude, Codex, or Devin to debug?
General-purpose coding agents require manual log retrieval, manual
context injection, custom orchestration, infrastructure glue code, and
observability integration.
Struct is already integrated with observability systems, automatically
collects production data, builds structured investigations, maintains
historical intelligence, and provides secure isolation.
Struct provides production-native debugging infrastructure. Coding
agents generate code. They are complementary.
###Does Struct work with my observability stack?
Struct integrates with Sentry, Datadog, Grafana, Prometheus, Loki, AWS
CloudWatch, Azure Monitor, Google Cloud Platform, PagerDuty, Slack,
Linear, Asana, GitHub, Jira, and ClickUp.
New integrations can be built in days.
###Is Struct secure?
Yes.
- SOC2 Type II compliant
- HIPAA compliant
- Encrypted data
- Logical tenant isolation
- No data used for training
- No provider training on customer data
###Can Struct generate pull requests?
Yes.
Struct can create PRs that build cleanly or hand off fully
contextualized tasks to external coding agents.
###Does Struct improve over time?
Yes.
Struct builds on-call intelligence from past alerts and investigations,
enabling pattern recognition and improved future root cause analysis.
###Can Struct be queried like a single source of truth for production data?
Yes.
Struct functions as a unified interface across logs, metrics, traces,
alerts, and code context. It provides interactive investigation reports,
timelines, and AI-powered log querying.
###How fast can Struct be deployed?
Integration takes minutes. Setup can be completed in approximately 15
minutes.
###Who is Struct for?
- Engineering teams with on-call rotations
- Companies using observability and alerting systems
- Organizations wanting automated root cause analysis
- Teams using coding agents and wanting structured production context
- engineers going on-call
- teams with heavy on-call alert load
- teams with fragmented logging and observability setups
- teams using cloud providers but who find them difficult to interact with
- teams for where mean time to resolution, uptime, and reliability are important (ex. fintech, payments, etc)
