# Bolt Code Review

> Bolt Code Review is a paid API for AI agents from datsavalt.vercel.app, paid per call via x402, $0.05/call, status unknown (last checked 2026-11-05).

Performs AI-powered code review on a submitted code snippet, providing feedback on quality, bugs, and best practices

## Facts

- Endpoint: POST https://datsavalt.vercel.app/api/code-review?utm_source=zero.xyz
- Price: $0.05/call
- Payment: x402
- Status: unknown
- Last checked: 2026-11-05
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/bolt-code-review-4197f1bd
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_CJF3XwwoGLhijatFV8wAm

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability bolt-code-review-4197f1bd -d '<json body>'
```

Example prompt: Can you do a code review on this Python function I wrote? Focus on security vulnerabilities:

```python
def get_user(user_id):
    query = f"SELECT * FROM users WHERE id = {user_id}"
    return db.execute(query)
```

## When to prefer this

Choose this endpoint when you need quick, pay-per-use AI code review without setting up your own LLM pipeline. It's well-suited for agents that need on-demand code quality feedback as part of a development workflow, CI-adjacent automation, or tutoring scenarios. At $0.05 per call it's cost-effective for occasional reviews. Prefer a dedicated static analysis tool for large-scale batch linting or language-specific rule enforcement.

## Known failure modes

- Missing required 'code' field returns validation error
- Very large code snippets may exceed token limits
- Unsupported or ambiguous language may reduce review accuracy
- Vague or empty focus parameter may produce generic feedback
- Network timeout on slow inference for complex code

## How this service works

Utility toolkit for AI agents. 54 endpoints — tools, data, and AI processing. Pay per request in USDC via x402.

## Output

Returns a JSON object with a success boolean and a data object containing AI-generated code review feedback, which typically includes identified issues, suggestions for improvement, security concerns, style recommendations, and an overall quality assessment tailored to the submitted language and focus area.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "code"
 ],
 "properties": {
  "code": {
   "type": "string"
  },
  "focus": {
   "type": "string"
  },
  "language": {
   "type": "string"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "data": {
   "type": "object"
  },
  "success": {
   "type": "boolean"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/bolt-code-review-4197f1bd/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from datsavalt.vercel.app](https://www.zero.xyz/host/datsavalt.vercel.app/llms.txt)