Skip to content

FAQ.md

Latest commit

 

History

History
55 lines (37 loc) · 2.92 KB

FAQ.md

File metadata and controls

55 lines (37 loc) · 2.92 KB

FAQ

General

What is CodeAgora?

A CLI tool where multiple AI models collaboratively review your code. Reviewers run in parallel, debate conflicting opinions, and a head agent makes the final verdict (ACCEPT / REJECT / NEEDS_HUMAN).

Is it free?

CodeAgora itself is free and open-source (MIT). You need API keys for LLM providers. Groq offers a free tier that works out-of-the-box.

Which providers are supported?

Curated API providers including OpenAI, Anthropic, Groq, OpenRouter, OpenCode Go, and OpenCode Zen, plus local CLI tools such as Claude Code, Codex, Gemini, Antigravity, OpenCode, Pi, Copilot, and Cursor. See PROVIDERS.md for the full list.

Reviews

What's the difference between ACCEPT, REJECT, and NEEDS_HUMAN?

  • ACCEPT: No critical issues found. Safe to merge.
  • REJECT: High-confidence critical issues detected. Fix before merging.
  • NEEDS_HUMAN: Low-confidence critical findings or unresolved debates. A human should verify.

Can I customize which models review my code?

Yes. Edit .ca/config.json or run agora init to select providers/models. You can also use --provider and --model flags per-review.

What are specialist personas?

Built-in review focus areas: builtin:security (OWASP), builtin:logic (race conditions, null checks), builtin:api-contract (breaking changes), builtin:general (maintainability). Assign them to reviewers in your config.

What is the "head" agent?

The L3 head agent makes the final ACCEPT/REJECT/NEEDS_HUMAN verdict based on all reviewer findings and debate outcomes. If the LLM-based head fails, a rule-based fallback is used automatically.

How do I see past reviews?

  • CLI: agora sessions to list, agora sessions show <date>/<id> for details
  • Desktop app: use the official cross-platform Tauri app for session browsing, result inspection, local review launch, and export workflows
  • Replay: agora replay <date>/<id> to re-render a past review

Configuration

Where is the config file?

.ca/config.json, .ca/config.yaml, or .ca/config.yml in your project root. Created by agora init.

Can I use YAML instead of JSON?

Yes. CodeAgora supports .ca/config.json, .ca/config.yaml, and .ca/config.yml. If multiple exist, precedence is config.json > config.yaml > config.yml.

How do I change the number of reviewers?

In config: set reviewers.count. Or per-review: agora review --reviewers 5.

Integration

Can I use CodeAgora with GitHub Enterprise?

Yes, set GITHUB_TOKEN for your GHE instance. The PR diff fetching uses gh CLI which supports GHE.

Can I use it in a monorepo?

Yes. CodeAgora detects project context (package.json, tsconfig.json) to reduce false positives across package boundaries.

Does it work with non-TypeScript projects?

Yes. The review pipeline works with any language. TypeScript diagnostics (pre-analysis) are skipped for non-TS projects automatically.