# The right model for every review, with costs you can see

MergeHarbor runs every review on its own models, picks the right one for each step, and shows what each review cost. No model keys to supply or pay for.

[Start free trial](https://app.mergeharbor.com/signup) [Read the docs](https://mergeharbor.com/docs/review-costs)

.mergeharbor.yml

```
review:
  tier: standard
  tier_rules:
    - name: Payments
      conditions:
        - { field: path, op: include, values: ["payments/**"] }
      tier: max
    - name: Large changes
      conditions:
        - { field: files_changed, op: gt, values: ["30"] }
      tier: deep
```

## How it works

1.  ### Choose a review tier

    Standard, Deep or Max, with rules by path or size of change in .mergeharbor.yml. MergeHarbor picks the models for each tier.

2.  ### Set budgets

    Set a monthly credit limit and a per-pull-request cap. When a review would go over, it steps down to a tier that fits.

3.  ### Track what reviews cost

    Each review run shows its cost per step, with the model, tokens and price list used. Cost per review shows the trend over time.

## What you get

Models included in your plan

Reviews and every other AI feature run on MergeHarbor's model accounts. Your plan's credits cover them, so there is no second bill from a model provider.

Failover between providers

If a model provider is unavailable, work moves to another provider your routing settings allow, so reviews still finish.

Data residency respected

Where your plan includes data residency, restrict model processing to approved providers and regions. Failover never moves work outside them.

The right model for each step

A small model checks risk and verifies findings. A stronger model runs only for risky changes and findings the small model could not settle.

Costs kept low by design

Prompt caching, leaving generated and vendored files out of prompts, and batch processing for work that can wait.

Alerts before you overspend

Billing admins get alerts at 80% and 100% of the allowance, and see credits per day by tier, developer and repository.

Documentation: [Review costs](https://mergeharbor.com/docs/review-costs) [Review tiers and credits](https://mergeharbor.com/docs/review-tiers) [How MergeHarbor uses AI and your code](https://mergeharbor.com/docs/ai-data)

## Frequently asked questions

Do we need our own model provider accounts or API keys?

No. The hosted service runs every review on MergeHarbor's own models. Your plan's credits cover model usage, and you never share or pay for model keys. Self-hosted Enterprise installs can use their own model provider accounts instead.

Which models review our code?

MergeHarbor routes each step to the model that fits it, from the providers your routing and data residency settings allow. Each review's cost breakdown names the model every step used.

What happens when a review would go over budget?

It steps down to a tier that fits the monthly credit limit and the per-pull-request cap.

Have another question? [Ask our support team](https://mergeharbor.com/contact?topic=support)

## Explore other features

### Codebase-aware review

Reviews that read your codebase, not just the diff

[Codebase-aware review](https://mergeharbor.com/features/review-agent)

### Free security checks

Built-in checks on every pull request, at no cost

[Free security checks](https://mergeharbor.com/features/security-review)

### CLI and local reviews

Review your changes before you push

[CLI and local reviews](https://mergeharbor.com/features/cli)

## Review every pull request with the context of your whole codebase

Connect a code host in minutes. Reviews start right away, and codebase context is added as indexing completes.

[Start free trial](https://app.mergeharbor.com/signup) [Book a demo](https://mergeharbor.com/contact?topic=sales&source=demo)