About Opsrift
AI-powered incident management: postmortems, shift handovers, runbooks, investigation, and status communication — built for the teams that keep systems running.
The Problem Opsrift Solves
Writing a thorough incident postmortem takes 1–3 hours. You open a blank document, switch between PagerDuty timelines, Datadog dashboards, Grafana alerts, Jira tickets, and Slack threads, and try to reconstruct what happened, why, and what to do about it. Most teams skip this step entirely — or produce inconsistent, shallow reports that don't drive improvement.
Shift handovers are worse. The outgoing engineer gives a verbal summary to the incoming one, and critical context gets lost. There's no structured record of what happened during the shift, what's still in progress, and what to watch out for.
Runbooks — the operational playbooks that should exist for every recurring incident type — rarely get created. The knowledge stays locked in the heads of senior engineers until they leave the team.
Customer-facing status updates are another gap: manual emails, fragmented social posts, and no single place for stakeholders to see component health and maintenance.
Opsrift addresses the full arc: import incident data from PagerDuty, OpsGenie, or Jira; generate postmortems, handovers, and runbooks in seconds; push work to Jira and Confluence; and (on the roadmap) investigate with AI and publish status pages — so every incident is documented and distributed, not just closed in a ticketing tool.
Why Opsrift Exists
Opsrift was built by Giga Kovaliovi, a senior operations professional with 6+ years of experience in iGaming and FinTech. Giga has led NOC and SOC teams handling 2,000+ monthly tickets, maintained 99.95%+ SLA across critical platform services, and reduced mean time to resolution (MTTR) from 17 minutes to under 4 minutes.
The frustration was always the same: the incident is resolved, but the documentation takes longer than the fix. Teams would skip postmortems, or write them weeks later from memory. Shift handovers were informal chats. Runbooks existed in one person's head.
Opsrift isn't a general-purpose AI writing tool. It's purpose-built for operations teams — with domain-specific prompts, native integrations with the tools ops teams already use, and output formats that match what your organization expects.
Key Capabilities
Postmortem Generation
Generate complete, 9-section incident postmortems with executive summary, timeline, root cause analysis, impact assessment, detection & response, resolution steps, contributing factors, action items, and lessons learned. Includes floating Table of Contents, mini-TOC scroll tracking, and appendix.
Shift Handover Generation
Create structured 7-section shift handover reports. Auto-fetch incidents from PagerDuty and Jira filtered by your shift window. Use quick range presets or custom windows. Shared integrations across all tools.
Runbook Generation
Build operational runbooks with 8 sections: overview, prerequisites, diagnostic steps, resolution procedures, rollback plan, escalation path, verification, and appendix. Create from scratch, import from PD/Jira, or generate directly from a completed postmortem.
Incident Investigation
AI-assisted investigation: ask natural-language questions about imported incidents, timelines, and context — without tab-hopping. Delivers structured analysis and recommendations grounded in your data.
Status Page
Hosted public status pages for customers: component health, incidents, maintenance windows, and optional branding. Central place for status communication instead of one-off emails and social threads.
Incident Forecast
Analyze your saved postmortem history to surface proactive risk signals: recurrence patterns across systems, deploy-to-incident correlation, unresolved action item debt, severity trend, and mean time between incidents (MTBI). An AI risk brief gives you an overall risk assessment so you can act before the next incident hits.
PagerDuty Integration
Import incidents with full timeline data, responders, affected services, and log entries. Filter and select from recent incidents. Shared across Postmortem, Handover, and Runbook generators.
OpsGenie Integration
Import alerts and incident data from OpsGenie for the same postmortem and handover workflows you use with PagerDuty — ideal for Atlassian-centric operations teams.
Jira Integration
Import issues by project, JQL filter, or keyword search. Push generated action items back to Jira as tickets with automatic priority mapping (Critical → Highest, High → High, etc.). Add cross-linked comments after Confluence publishing.
Confluence Publishing
One-click publishing converts Markdown to Atlassian Document Format. Tables, code blocks, blockquotes, and inline formatting are preserved. Jira ticket IDs are automatically hyperlinked to their Jira URLs.
Slack & Microsoft Teams
Webhook notifications for postmortem summaries and status updates so leadership and adjacent teams see outcomes in the channels they already monitor — Slack via Block Kit, Teams via incoming webhooks.
GitHub Deploy Correlation
Correlate deployments with incident windows to speed root-cause analysis and timeline reconstruction without manual git archaeology.
Inline Editing & Regeneration
Click any section heading to enter edit mode. Make manual changes or regenerate a single section with AI while keeping the rest intact. Changes reflect immediately in the full report.
Export & Sharing
Export finished reports as PDF, copy as Markdown, or format for Slack. Save reports to your dashboard for future reference. Share via Confluence or direct copy.
Use Cases
Post-Incident Documentation
After resolving an incident, import it from PagerDuty or Jira and generate a complete postmortem in under 60 seconds. Edit the AI output, push action items to Jira, and publish to Confluence — all from one screen.
Shift-to-Shift Handover
At the end of every shift, generate a structured handover report that captures active incidents, ongoing tasks, escalations, key metrics, and handover notes. The incoming engineer gets full context without a 30-minute verbal briefing.
Runbook Library Building
After every major incident, generate a runbook from the postmortem. Over time, build a searchable library of operational runbooks that capture institutional knowledge and reduce MTTR for recurring issues.
Customer Status Communication
Publish component status, incidents, and maintenance to a hosted status page so customers have a single source of truth — reducing duplicate email and ad-hoc social updates during outages.
AI-Assisted Incident Investigation
During or after an incident, query structured incident context with natural language to explore timelines, impact, and follow-ups — complementing your runbooks and dashboards without replacing them.
Proactive Risk Forecasting
Before an incident happens, use Incident Forecast to review risk signals across your postmortem history: which systems recur most, whether your deploy cadence correlates with outages, whether past action items are aging unresolved, and whether severity is trending worse. Catch the pattern before it becomes a P1.
Compliance & Audit Trail
Standardized, timestamped incident reports meet the documentation requirements of SOC 2, ISO 27001, and industry-specific regulations. Every report follows the same structure, every time.
Team Standardization
Eliminate the 'blank page' problem. Whether an engineer has written 100 postmortems or zero, the output follows the same 9-section structure with consistent depth and actionable recommendations.
How Opsrift Works Under the Hood
Opsrift uses Anthropic's Claude API (Claude Sonnet) for all AI generation. Each section of a report is generated with a specialized prompt by a dedicated API call, running in parallel for speed. A 9-section postmortem is typically generated in 15–30 seconds.
The frontend is built with React and TypeScript, using Tailwind CSS for the cyber-noir design system. The backend is an Express.js API deployed on Vercel. Authentication is handled by Clerk, data storage by Supabase with Row Level Security, and payments by Paddle.
Your incident data is never used to train AI models. It is transmitted securely via HTTPS, processed by Claude for generation, and stored in your Supabase account. Opsrift does not retain generated content beyond your authenticated session and saved reports.
Start Your Free Trial
7-day free trial with full access. Card required, cancel anytime.