Frequently Asked Questions

General

1. What is xhr.dev?

xhr.dev is an on-prem anti-bot challenge solver. It ships as a Docker container you run inside your own infrastructure, and it solves Akamai Bot Manager and DataDome challenges via a local HTTP/WebSocket API.

2. Is there a hosted / cloud version?

No — xhr.dev is on-prem only. You run the container yourself; nothing proxies through xhr.dev's infrastructure. See Deployment.

3. What does xhr.dev solve today?

• Akamai Bot Manager — _abck / bm-sz sensor challenges and SBSD
• DataDome — captcha and interstitial challenges

Don't see a provider you need? Message us.

4. How do I integrate it?

Everything is plain HTTP/WebSocket to your own running container. The choice is whether you need a browser:

1. Without a browser — your code makes the requests and hands the captured challenge to POST /akamai/solve or POST /dd/solve. See Via HTTP (Node, Python), or Via the API for the raw flow in curl.
2. With a browser — when the site needs one anyway; Akamai also has a WebSocket session (/akamai/session) for this. See Playwright or Lightpanda.

Start at How to integrate, or go straight to the API references for Akamai and DataDome.

If you'd rather have a coding agent do it, Via Claude has a prompt and a skill you can install.

5. How is it licensed?

Per-customer, via a signed licence file (licence.json + licence.sig) that the container verifies locally at startup — no phone-home to xhr.dev. Contact us to get one issued. See Deployment for how licence delivery and renewal work.

6. What does it cost?

Contact us — pricing depends on scale and which providers/features you need.

7. What are the system requirements?

Docker, on either ARM64 or x86-64. No GPU or special hardware required.

8. Does xhr.dev see my traffic, proxies, or target data?

No. The container runs entirely on your infrastructure. There's no outbound call to xhr.dev at runtime — the licence check is a local signature verification, not a network request. See Security.

9. Can I customize its behavior?

Yes — per-request fields (browser profile, proxy, timeouts, submit vs. capture-only mode) are documented in the Akamai and DataDome references. Reach out if you need something the current API doesn't expose.

10. How do I get support?

Book a call, email us, or find us on GitHub and LinkedIn.

11. How often is the image updated?

xhr.dev ships updated images as challenge scripts change upstream. Because the licence is decoupled from the image, you can usually pull a new image tag without any licence changes; if a licence renewal is due separately, you'll hear from us ahead of expiry.

12. Am I responsible for keeping it up to date?

Yes for self-hosted deployments — you control when you pull a new image tag. xhr.dev publishes updated images as anti-bot vendors change their challenges; we'll flag it when a change means you should update sooner rather than later.

13. Can it handle multiple bots / high concurrency?

Yes. Each container admits solves through an internal queue — 8 concurrent solves by default — and you scale horizontally by running more containers behind a load balancer.

14. Are there usage limits or quotas?

None enforced by the software itself — throughput is bounded by your own hardware and the licence terms you agreed to (e.g. seat/host count, if applicable). Contact us if you're unsure what your licence covers.

15. What level of coding knowledge is required?

You need to be comfortable making HTTP/WebSocket requests and running Docker. No SDK is required to get started — see the curl examples in Getting Started.

16. Does it work with any programming language?

Yes — it's a plain HTTP/WebSocket API, callable from any language that can make HTTP requests.

17. What kind of monitoring or logging does the container produce?

It logs locally (stdout/your log driver of choice) — nothing is sent to xhr.dev. See Security for details on what's logged.

18. Is there authentication on the API?

No, and there's no API key to obtain. The licence gates startup, not requests, so anything that can reach port 3000 can use the solver. That's by design — the container is meant to run on a locked-down host reachable only from your own scrapers. See Network posture.

(The x-api-key you'll see in the examples repo is for the hosted trial box we lend out during evaluations, which sits behind a reverse proxy. Your own deployment ignores it.)

19. Can I have an AI agent write the integration?

Yes, and it's a good fit — the job is well-specified and there are runnable examples to copy. Via Claude has a prompt to paste, a Claude Code skill to install, and the list of things agents reliably get wrong here. The docs are also published as llms.txt / llms-full.txt so an agent can read the whole site in one fetch.