OpenShrike - agentic code reviewer vs ReviewPhin

Self-hosted agentic code review for GitHub/GitLab, but focuses on integration rather than Markdown-driven policies.

Side by side

What it is: OpenShrike - agentic code reviewer:OpenShrike enforces engineering standards as Markdown-driven checks that run automatically on every PR.; ReviewPhin:Self-hosted, agentic code review tool that automates GitLab and GitHub pull-request reviews on your own infrastructure.
Best for: OpenShrike - agentic code reviewer:Markdown-driven automated code review; ReviewPhin:Agentic code review on self-hosted infra
Who it’s for: OpenShrike - agentic code reviewer:Teams that need automated, policy-based code reviews for AI-generated or regular code.; ReviewPhin:DevOps operators and engineering teams that want automated code reviews without losing deployment control.
Pricing: OpenShrike - agentic code reviewer:Open source; ReviewPhin:Free
Open source: OpenShrike - agentic code reviewer:Yes, 5 GitHub stars; ReviewPhin:Yes, 16 GitHub stars
Works with: OpenShrike - agentic code reviewer:OpenAI, Anthropic, LMStudio, Ollama, Zen, Bedrock, Azure; ReviewPhin:GitLab, GitHub, Anthropic, Docker Compose
DevHunt upvotes: OpenShrike - agentic code reviewer:5; ReviewPhin:12
Launched on DevHunt: OpenShrike - agentic code reviewer:Jun 2026; ReviewPhin:Jul 2026

Markdown-driven checks. Define architectural and security requirements in Markdown files and have them enforced on each PR.
Built-in best-practice library. Includes 220+ checks across 32 language policies to bootstrap review standards.
Auto-fix with agents. Spawns an LLM agent to automatically fix failing checks before human review.
Self-hosted & BYOK. Runs locally or in CI and lets you bring your own LLM key (OpenAI, Anthropic, Ollama, etc.).
VS Code extension. Run scans, view findings, and apply fixes directly from the editor without leaving your machine.
CLI tooling. Commands like shrike init, scan, and fix enable quick setup and automated enforcement.

One-click deployment. Run ReviewPhin via Docker Compose or Helm chart; a single container hosts the worker, setup flow, and docs.
Platform registration. Store reusable GitLab or GitHub credentials as platform connections and attach them to projects.
Model flexibility. Use bundled Copilot CLI or configure OpenAI-compatible, Azure, or Anthropic endpoints via named profiles.
Pluggable storage. Start with SQLite, switch to Flotiq, or add a custom storage adapter for review persistence.
Automated review publishing. Trigger reviews from merge-request surfaces; findings, replies, and summaries are posted as native bot-owned comments.
Container-ready runtime. The same image serves the worker, setup assets, and documentation, simplifying production deployment.

Based on each tool's website and DevHunt data. Details may change; check the official sites.