Proval

Proval

Self-hosted LLM code review agent that automatically reviews pull requests and replies to issues across GitHub, GitLab, and other Git platforms.

🔗 Visit Proval
📁 Editors, IDEs & Dev Tools🗣️ English📅 August 30, 2026

Description

Code review tools that use AI usually mean sending your source code to a third-party cloud service, which is a non-starter for teams with strict privacy requirements or simply a preference for keeping everything in-house. Proval is built for exactly that situation: it's a code review bot you run yourself, on your own servers, using either your own API keys or a fully local language model.

Proval integrates with GitHub, GitLab, Forgejo, Gitea, and Codeberg to automatically review pull requests, grouping files and using specialist sub-agents per group, leave inline comments on affected lines, and reply to issue threads. It supports any OpenAI-compatible API as well as fully local models via Ollama, tracks token usage for cost visibility, and runs on a self-hosted stack (Bun, Hono, SvelteKit, SQLite) deployable via Docker or Kubernetes.

💬 Our review

The short version: Proval is a solid pick specifically for teams that want AI code review but can't or won't send code to a third-party cloud service, the self-hosted, bring-your-own-model design is the whole point.

Cloud-based competitors like CodeRabbit or Sourcery offer more polished, zero-setup experiences and continuously updated review quality, but they require sending your code off-premises; Proval trades some of that polish for full control over where your code and review data live, including the option to use a fully local Ollama model with zero external calls at all. Support for lesser-covered platforms like Forgejo, Gitea, and Codeberg is a genuine plus for self-hosted Git users. Worth adopting if privacy or self-hosting is a hard requirement; if it isn't, a managed cloud reviewer will likely need less setup and tuning.

💰 Pricing

GratuitGratuit, AGPL-3.0 — coût de l'API LLM ou de l'infra locale à part
Open-source $0 (+ coût de votre API/modèle)

📊 Global score

53Average
🌐Availability15/100Faible

1 language · 0 platform

📄Profile90/100Excellent

Profile completeness

🤖 AI-enriched data

💰 Pricing model
🆓 Gratuit

Gratuit, AGPL-3.0 — coût de l'API/infra à part

👥 Target audienceÉquipes dev voulant une revue de code IA sans envoyer leur code à un tiers
🗣️ Languagesen
🌍 Target countriesInternational
👍

Pros

auto-hébergé — le code ne part jamais vers un tiers

compatible modèles locaux (Ollama), zéro appel externe possible

couvre GitHub, GitLab, Forgejo, Gitea, Codeberg

sous-agents spécialisés par groupe de fichiers

gratuit, open-source (AGPL-3.0)

👎

Cons

nécessite d'héberger et maintenir soi-même l'infrastructure

moins polie/prête à l'emploi qu'une solution cloud managée

qualité de revue dépendante du modèle choisi (local vs cloud)

❓ Frequently asked questions

What is Proval?
Is Proval free?
Can Proval run with a fully local LLM?
Which Git platforms does it support?
Do I need to host it myself?
Is it worth the money compared to alternatives?
Which tool should you pick for your case?