LiveReview

What LiveReview is

LiveReview is a risk-aware AI code review platform from Hexmos. It scores every PR/MR and, under your policy, auto-approves the lowest-risk 30-50% of changes so reviewers only spend time on risky ones. It works on code from any author: people, Claude, Copilot, Cursor or other agents.

How it works

• Risk score: blends semantic LLM analysis, a custom ML model trained on your incidents and reverts, and static call-graph analysis (inspired by Meta's RADAR research).
• Policies you can test: conservative, balanced or aggressive; replay on past PRs first; force-approve with a reason; calibration report with false-approve rate.
• Blast radius and review priority: every hunk is ranked by reach, persistent-state writes, complexity and test coverage.
• 10 risk categories, 100+ failure patterns across reliability, correctness, performance, scalability and security.

Where it runs

• GitHub App plus PR threads on GitHub, GitLab, Bitbucket, Gitea and Azure DevOps
• git-lrc commit-time reviews, claude-lrc for Claude Code, VS Code/Cursor/Antigravity extension
• Slack, Teams, Discord, MCP and REST API; scheduled reviews of the default branch
• Own fine-tuned model, or bring keys for Gemini, OpenAI, Claude, Bedrock, Ollama

Pricing

Free: 30k reviewed LOC/month, up to 3 members. Premium: from $32/month for 100k LOC with unlimited users, scaling to $1,024 for 3.2M LOC. Enterprise: self-hosted Docker deployment, SSO, custom domain and SLA.

Company

Built by Hexmos Technology LLP in Bengaluru, India. Source-available on GitHub.