How much does open-code-review cost?

open-code-review itself is free and Apache-2.0 licensed. The only cost is the LLM tokens you spend on your own OpenAI or Anthropic key when a review runs. The maintainers report the tool uses about one-ninth the tokens of Claude Code on the same 200-pull-request benchmark.

Which models can I plug into open-code-review?

open-code-review talks to any OpenAI-compatible or Anthropic-compatible endpoint. The v1.7.16 release added iFlytek Spark as a built-in provider, and earlier 1.7 releases added LiteLLM and Eden AI gateways, so you can route to most hosted or self-hosted models without patching the tool.

How is open-code-review different from Claude Code review?

Claude Code is a general coding agent; open-code-review is a specialist. It fires a deterministic ruleset for NPE, thread-safety, XSS, and SQL injection first, then only spends LLM tokens where the pipeline flagged risk. On a 200 PR benchmark from 50 open-source repos, Alibaba reports higher precision and F1 than Claude Code at roughly 1/9 the token cost.

What is new in v1.8.0?

open-code-review v1.8.0, released on 2026-07-28, adds remote Model Context Protocol servers over Streamable HTTP, review support for Bicep and HCL/Terraform files, and chunking of inline comments into bounded batches. Fixes cover VSCode workspace review before the first commit, Go module path alignment and a js-yaml DoS upgrade.

Alibaba · 2026-07-28 · major

Alibaba open-code-review — line-level LLM code review at 1/9 the tokens

Alibaba's Apache-2.0 code-review CLI hit v1.8.0 on July 28 with remote MCP servers over Streamable HTTP, Bicep and Terraform review, and chunked inline comments. The Go tool now sits at 15.7k GitHub stars and pairs deterministic rules with an LLM agent.

GitHub social card for alibaba/open-code-review, an open-source AI code review CLI

Hybrid pipeline + LLM code reviewer, battle-tested at Alibaba, now Apache-2.0 with 15.7k GitHub stars.

Quick facts

Maker	Alibaba
License	Apache-2.0
Latest version	v1.8.0 (2026-07-28)
Language	Go
Providers	OpenAI + Anthropic compatible
Install	npm install -g @alibaba-group/open-code-review
GitHub stars	15.7k

What is it?

open-code-review is a command-line code reviewer that reads a Git diff and posts precise, line-anchored comments. Alibaba built it internally and released it under Apache-2.0. The project pairs deterministic rule-based pipelines with an LLM agent, and it plugs into OpenAI-compatible or Anthropic-compatible model providers.

How does it work?

Reviews start with the deterministic pipeline, which applies a fine-tuned ruleset for common defects — NPE, thread-safety, XSS, SQL injection — and picks which files or chunks are worth spending tokens on. The LLM agent then does dynamic context retrieval and writes the line-level comments. Version 1.8.0, shipped on 2026-07-28, added remote MCP over Streamable HTTP, Bicep and HCL/Terraform review, and batching of inline comments into bounded chunks so large PRs no longer overrun the review UI.

Why does it matter?

Teams that already run Claude Code or Copilot review can replace them without changing model providers, since open-code-review speaks OpenAI and Anthropic APIs. The maintainers benchmark it on 200 real pull requests from 50 open-source repos and report higher precision and F1 than Claude Code at about 1/9 the token cost — the kind of gap that shows up on a monthly invoice.

Who is it for?

platform teams and open-source maintainers

Frequently asked questions

How much does open-code-review cost?: open-code-review itself is free and Apache-2.0 licensed. The only cost is the LLM tokens you spend on your own OpenAI or Anthropic key when a review runs. The maintainers report the tool uses about one-ninth the tokens of Claude Code on the same 200-pull-request benchmark.
Which models can I plug into open-code-review?: open-code-review talks to any OpenAI-compatible or Anthropic-compatible endpoint. The v1.7.16 release added iFlytek Spark as a built-in provider, and earlier 1.7 releases added LiteLLM and Eden AI gateways, so you can route to most hosted or self-hosted models without patching the tool.
How is open-code-review different from Claude Code review?: Claude Code is a general coding agent; open-code-review is a specialist. It fires a deterministic ruleset for NPE, thread-safety, XSS, and SQL injection first, then only spends LLM tokens where the pipeline flagged risk. On a 200 PR benchmark from 50 open-source repos, Alibaba reports higher precision and F1 than Claude Code at roughly 1/9 the token cost.
What is new in v1.8.0?: open-code-review v1.8.0, released on 2026-07-28, adds remote Model Context Protocol servers over Streamable HTTP, review support for Bicep and HCL/Terraform files, and chunking of inline comments into bounded batches. Fixes cover VSCode workspace review before the first commit, Go module path alignment and a js-yaml DoS upgrade.

Try it

npm install -g @alibaba-group/open-code-review