Open Code Review
Coding assistants and agents · Alibaba · Open source project
Hybrid architecture code review tool combining deterministic pipelines with LLM agents for precise line-level comments. open-codereview.ai ↗
Facts
DescriptionOpen Code Review is a CLI-installable AI agent that performs line-level code review comments on files, branches, diffs, and commits. It combines deterministic engineering modules with LLM agent reasoning, supports multiple LLM providers and custom endpoints, and claims full data privacy since reviews run against the user's chosen model. It also publishes an open benchmark (AACR-BENCH) comparing models on real pull requests.
DeploymentOpen sourceAPI
PricingFree
Target buyersengineering
IntegrationsNpmHomebrewAnthropic Messages APIOpenAI Chat Completions APIOpenAI Responses APIDashScopeDeepSeekZ.AI
Compliance claimsNo training on customer data
Models usedAnthropic ClaudeOpenAIDashScope (Qwen)DeepSeekZ.AI (GLM)
Notable customersAlibaba Group (internal use claimed)
Features · 14
line-level review commentsbranch and diff reviewcommit reviewmulti-model supportcustom llm endpointseffort-driven review depthmemory compressionhallucination reflection moduleprecise comment positioningbuilt-in review rules40+ language coveragesecurity checks sql injection xssbenchmark publishingcli install
Dossier · model-written summary
Skeptical read: this is a genuine AI tool, a terminal-based code review agent open-sourced by Alibaba. Metrics on the page render as 0 (0K+ users, 0M+ tasks), suggesting broken stat display, so adoption claims cannot be verified. Benchmark results are self-published and favor Alibaba models. No stated pricing (free, open source), no compliance certifications, and 'keep data fully private' depends
Named customers: Alibaba Group (internal use claimed)
Timeline · 1 event
Same category · Coding assistants and agents