> ## Documentation Index
> Fetch the complete documentation index at: https://mintlify.com/speedyapply/JobSpy/llms.txt
> Use this file to discover all available pages before exploring further.

# Using proxies

> Configure proxies to avoid rate limiting when scraping job boards, especially LinkedIn.

Job board websites rate-limit requests from the same IP address. When you exceed their threshold, you receive a `429 Too Many Requests` response and scraping stops. Proxies let you rotate IP addresses to reduce this risk.

## Which boards need proxies

| Board        | Rate limiting      | Recommendation       |
| ------------ | ------------------ | -------------------- |
| Indeed       | Minimal            | Not required         |
| ZipRecruiter | Moderate           | Optional             |
| Glassdoor    | Moderate           | Optional             |
| LinkedIn     | Highly restrictive | Strongly recommended |

<Note>
  LinkedIn typically rate-limits around the 10th page of results from a single IP address. If you need more than \~100 LinkedIn results per run, proxies are effectively required.
</Note>

## The `proxies` parameter

Pass a list of proxy strings to `scrape_jobs()` via the `proxies` parameter. Each scraper rotates through the list in round-robin order.

```python theme={null}
from jobspy import scrape_jobs

jobs = scrape_jobs(
    site_name=["linkedin", "indeed"],
    search_term="software engineer",
    location="New York, NY",
    results_wanted=50,
    proxies=[
        "user:pass@host1:port",
        "user:pass@host2:port",
        "localhost",  # fall back to direct connection
    ],
)
```

### Proxy format

Proxies are strings in the format `user:pass@host:port`. You can also use `"localhost"` to represent a direct (no-proxy) connection slot in the rotation.

```python theme={null}
proxies = [
    "alice:secret@192.168.1.10:8080",
    "bob:pass123@10.0.0.5:3128",
    "localhost",
]
```

JobSpy also accepts URLs with explicit schemes:

```python theme={null}
proxies = [
    "http://user:pass@proxy.example.com:8080",
    "https://user:pass@proxy.example.com:8443",
    "socks5://user:pass@proxy.example.com:1080",
]
```

### Single proxy

You can pass a single proxy as a string instead of a list:

```python theme={null}
jobs = scrape_jobs(
    site_name="linkedin",
    search_term="backend engineer",
    location="San Francisco, CA",
    proxies="user:pass@proxy.example.com:8080",
)
```

## How rotation works

Each scraper instance gets its own rotating proxy session. When a scraper makes a request, it advances to the next proxy in the cycle. If one proxy is blocked, the next request will use a different proxy automatically.

This means if you are scraping four sites simultaneously, each site rotates through the proxy list independently.

## CA certificate for proxies

Some corporate or SSL-intercepting proxies require a custom CA certificate for HTTPS inspection. Pass the path to the certificate file via `ca_cert`.

```python theme={null}
jobs = scrape_jobs(
    site_name=["linkedin", "indeed"],
    search_term="devops engineer",
    location="Austin, TX",
    proxies=["user:pass@corporate-proxy.example.com:8080"],
    ca_cert="/path/to/ca-bundle.crt",
)
```

## Overriding the user agent

The default user agent string may become outdated as job boards update their bot detection. Use `user_agent` to override it with a current browser user agent.

```python theme={null}
jobs = scrape_jobs(
    site_name="linkedin",
    search_term="machine learning engineer",
    location="Seattle, WA",
    user_agent=(
        "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) "
        "AppleWebKit/537.36 (KHTML, like Gecko) "
        "Chrome/124.0.0.0 Safari/537.36"
    ),
)
```

## Handling rate limit errors

A `429` response means the job board has temporarily blocked your IP. When this happens:

1. Wait before scraping again. The required wait time is site-dependent — LinkedIn may need several minutes, while other boards recover faster.
2. Add more proxies to the rotation to distribute requests across more IP addresses.
3. Reduce `results_wanted` to make fewer requests per run.

<Warning>
  All job board endpoints are aggressive with blocking. Never scrape in a tight loop without delays.
</Warning>