Free
适合个人开发者
$0
- 无限代码库理解
- 每月 20 次智能体修复
- 使用自己的模型密钥
- 本地代码库 + 私有代码库
- 永久自托管完整版
包含内容
每个方案均从理解开始。
Why it's priced this way: understanding your repo is deterministic and costs almost nothing — so triage is unlimited on every plan. Agent fixes spend model tokens, so you pay only when you scale.
| Free | Pro | Team | Enterprise | |
|---|---|---|---|---|
代码库理解 triage | 无限 | 无限 | 无限 | 无限 |
智能体修复 fix | 20 / mo | High | Org pool | Unlimited |
| LLM | BYOK | Managed | Managed | Managed / on-prem |
| GitHub App + 自动修复 PR | — | ✓ | ✓ | ✓ |
| 并发运行 | 1 | Multiple | Multiple | Multiple |
| 席位 / SSO | — | — | ✓ | ✓ |
| 自托管 | ✓ forever | ✓ | ✓ | ✓ + support |
| 审计 / 治理 / SLA | — | — | — | ✓ |
Every fix is verified by your own test suite before it lands.
Frequently asked
Triage is unlimited at no cost. You can map repositories, explore the dependency graph, and see blast-radius analysis — all for zero tokens, zero spend.
Yes. The self-hosted tier is free forever. You bring your own key, your code never leaves your infrastructure, and there is no expiration.
You connect your own API key from your preferred provider. Codna charges nothing for inference — you pay your provider directly, typically pennies per fix.
Fix generation is metered on managed plans. Triage — mapping repos, surfacing evidence bundles, identifying blast radius — stays unlimited regardless of plan.
The deterministic engine reduces each fix to a ~600-token evidence bundle, measured at 162x less context than reading the repo. Less context means dramatically lower inference cost — 5× fewer tokens than Cursor and 1.7× faster in head-to-head tests.
Yes. Managed plans handle hosting and scale with your usage. You get the same verified, test-confirmed fixes — without running your own infrastructure.