Crawlora Rust SDK

Official async Rust client for Crawlora's public web data API. It includes generated endpoint helpers for the public API contract, typed request parameters and response models, API-key/JWT authentication, configurable timeouts, retry handling, per-request headers, and pagination helpers.

The current crate contract is generated from the public OpenAPI export in openapi/public.json. Each endpoint method returns the complete HTTP response so callers can deserialize it into a generated model or their own type.

Install

[dependencies]
crawlora = "1.49.0-sdk.1"
tokio = { version = "1", features = ["macros", "rt-multi-thread"] }

Requires Rust 1.88 or newer and an async Tokio runtime.

Quick start

use crawlora::{BingSearchParams, CrawloraClient};

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let api_key = std::env::var("CRAWLORA_API_KEY")?;
    let client = CrawloraClient::new(api_key)?;

    let response = client
        .bing
        .search(BingSearchParams {
            q: Some("coffee shops in Seattle".to_owned()),
            ..Default::default()
        })
        .await?;

    let json: serde_json::Value = response.into_json()?;
    println!("{}", serde_json::to_string_pretty(&json)?);
    Ok(())
}

Set CRAWLORA_API_KEY in the process environment. The key is sent in the x-api-key header and should stay on the server side.

Configuration

use std::time::Duration;
use crawlora::CrawloraClient;

let client = CrawloraClient::builder()
    .api_key(std::env::var("CRAWLORA_API_KEY")?)
    .base_url("https://api.crawlora.net/api/v1")
    .timeout(Duration::from_secs(45))
    .max_retries(3)
    .build()?;
# Ok::<(), Box<dyn std::error::Error>>(())

Use .jwt_token(...) for endpoints authenticated with a bearer token. RequestOptions can override headers, timeout, retry count, and response mode for an individual call.

Typed response models

Endpoint helpers return ApiResponse, which preserves the HTTP status and headers. Deserialize JSON into one of the generated models or your own type:

use crawlora::{BingSearchParams, CrawloraClient, ModelBingSearchResponse};

async fn search(client: &CrawloraClient) -> Result<ModelBingSearchResponseDoc, crawlora::Error> {
    client.bing.search(BingSearchParams {
        q: Some("coffee".to_owned()),
        ..Default::default()
    }).await?.into_json()
}

The dynamic client.call(operation_id, params) method is available when an operation is selected at runtime. See docs/operations.md for the complete generated endpoint list and docs/recipes.md for pagination and response-mode examples.

Errors and retries

Non-2xx responses return Error::Api with the HTTP status, response body, and request ID when provided by the API. Retries honor Retry-After and use bounded exponential backoff for retryable responses. Non-idempotent requests are not retried on server errors unless explicitly enabled in RequestOptions.

License

MIT. See LICENSE.