If you are scraping prices on Wildberries, managing 20 Instagram accounts, or running ads through Facebook Ads, sooner or later you will see one of three things: error 403, error 429, or captcha. Most beginners react the same way — they change the proxy to any other and hope it will work. In fact, each error code is a message from the site about a specific reason for the block, and it can accurately determine what type of proxy will solve the problem: data center, residential, or mobile.
Why the error code is more important than the fact of blocking
When a site blocks a request, it almost always returns an HTTP status code — this is a standard "label" that the browser or scraper receives along with the page. The problem is that many users of anti-detect browsers and scrapers look only at the fact of "not opening" and do not check what specific code was returned. And the difference is huge: 403 means that the site recognized you as an unwanted guest and intentionally closed access. 429 means that you have physically exceeded the request limit — the site is not against you personally, it just needs you not to hit the server so often. Captcha is an intermediate status: the site doubts whether you are a human or a bot and gives you a chance to prove otherwise.
Each of these situations requires a different solution. If you receive a 403 and simply change one data center proxy to another similar one — you will get 403 again because the reason is not in the specific IP, but in the category of IP. If you receive a 429 and buy expensive mobile proxies instead of reducing the frequency of requests — you will waste your budget because the issue is not with the type of IP, but with the speed of requests. Understanding the error code saves both time and money on proxies.
Error 403 Forbidden: the site recognized you and does not want to let you in
403 is an intentional refusal. The site received your request, processed it, and decided: this IP or this set of characteristics is not allowed access. Most often, 403 occurs for three reasons:
- IP from the "blacklist" of data centers. Sites like Wildberries, Ozon, Avito, and most social networks maintain databases of IP address ranges owned by cloud providers (AWS, Google Cloud, DigitalOcean, etc.). If your proxy is data center-based, and its subnet has already been "exposed" by other scrapers, you will receive 403 instantly, even on the first request.
- Geographical restriction (geo-blocking). Some content or advertising accounts are only available from certain countries. Accessing from an IP not from that region — this is also 403.
- Browser fingerprint does not match the expected one. If you are using a proxy in conjunction with an anti-detect browser (Dolphin Anty, AdsPower, Multilogin, GoLogin), and the browser fingerprint does not match the declared geolocation of the IP — for example, IP from Russia, but the timezone and system language are English — the site may block you based on the combination of characteristics.
The solution for persistent 403s: switch from data center proxies to residential proxies. These are IPs from real home providers — they are not listed in cloud ranges and appear to the site as a regular visitor from home. For social networks and advertising platforms, where 403 is accompanied by a request to confirm the device, often only mobile proxies save the day — they are tied to IPs from cellular operators and raise minimal suspicion even under aggressive checks.
Error 429 Too Many Requests: you are being throttled by speed, not by identity
429 is not about who you are, but about how often you knock on the server. The rate limit is triggered when more requests come from one IP or session than allowed in a given time frame. A typical scenario: you set up competitor price monitoring on Ozon and decided to check the entire catalog every 5 minutes with the same IP — the server will quickly impose a limit and start returning 429 instead of data.
A key mistake for beginners is to immediately change the type of proxy to a more expensive one when receiving a 429, although the problem can be solved differently:
- Reduce the frequency of requests. Add delays between requests (2-5 seconds for scraping marketplaces is usually sufficient).
- Rotate IPs more frequently. If the same IP makes 100 requests in a row, the limit will be reached regardless of the type of proxy. Here, rotation for each request or every N requests helps.
- Increase the pool of IPs. If the task is to maintain a high speed of data collection, you need more IP addresses to distribute the load, rather than a "higher quality" single IP.
For mass scraping, where speed is more important than anonymity (for example, collecting public prices without authorization), data center proxies are well-suited — they are faster and cheaper, and speed limits can be resolved by increasing the pool and proper rotation, rather than changing the type of IP.
Captcha: the site gives you a chance but suspects automation
Captcha is an intermediate status between "all is well" and "access denied." The site sees signs that resemble a bot: too regular intervals between actions, lack of mouse movements, atypical HTTP header sets, IP from a range that is often associated with automation. Instead of a hard 403, it gives you the opportunity to confirm that you are human.
Frequent captcha is a signal that the problem is not in blocking as such, but in behavioral patterns and the reputation of the IP. If captcha appears on Instagram or TikTok when registering new accounts, it is almost always due to the combination: data center IP + new account + abrupt actions without "warming up." In this case, switching to residential or mobile proxies significantly reduces the frequency of captcha because the reputation of such IPs is higher — real users access them, not just bots.
Another common case is captcha on marketplaces when scraping without authorization. Here, not only changing the type of proxy helps, but also simulating human behavior: random delays, scrolling before clicking, realistic User-Agent. Proxies alleviate some suspicion, but do not replace correct behavior of the script or scenario in an anti-detect browser.
Table: error code → probable cause → type of proxy
| Code / symptom | Probable cause | What to change | Recommended type of proxy |
|---|---|---|---|
| 403 Forbidden on the first request | IP in the database of data center ranges | Type of IP, not the number of requests | Residential proxies |
| 403 when logging into a social network / advertising account | Mismatch between fingerprint and geolocation of IP | Anti-detect browser settings + IP | Mobile proxies |
| 429 Too Many Requests | Request limit exceeded from one IP | Request frequency, rotation, pool size | Data center proxies (with rotation) |
| Captcha when registering new accounts | Low IP reputation + new account | IP reputation and "warming up" the account | Mobile proxies |
| Captcha during mass scraping without login | Bot behavioral patterns | Behavior simulation + IP change | Residential proxies |
Practical cases: Wildberries, Ozon, Instagram, Facebook Ads
Wildberries and Ozon: price monitoring
Sellers who monitor competitor prices most often encounter 429 rather than 403. Marketplaces limit request speeds at the API and web interface level. The right solution is not to chase after "elite" types of proxies, but to set up a pool of several hundred data center IPs with rotation for each request or every few requests, plus a delay of 1-3 seconds between requests for the same product. If after this captcha starts appearing — then it is worth connecting residential IPs, because captcha in this case signals a behavioral, not a speed issue.
Instagram and TikTok: farming and managing accounts
SMM specialists managing 20-30 client accounts through Dolphin Anty or AdsPower usually see 403 or a phone confirmation request immediately after linking a data center IP to a new profile. This is a classic sign that a switch to mobile proxies is needed — Instagram and TikTok are particularly sensitive to the type of IP when working with new or "cold" accounts.
Facebook Ads: farming advertising accounts
Arbitrageurs launching 10-20 advertising accounts often receive a block not as a clear 403 at the network level, but as a ban at the account level after a series of suspicious actions. Here, the error code is not always visible directly, but the logic is the same: if bans come in batches immediately after switching the IP to a new data center range — the problem lies in the type of proxy. Switching to residential or mobile IPs with a unique subnet for each account reduces cascade bans (when blocking one account drags down others linked to the same IP).
Diagnosis checklist before changing proxies
- Record the exact error code (403, 429, 503) or type of captcha — not just "didn't open."
- Check if the code repeats from the first request (a sign of 403 — type of IP) or after a series of requests (a sign of 429 — speed).
- Check the current type of proxy: data center, residential, or mobile.
- If 403 occurs on the first request — switch to residential or mobile IPs.
- If 429 — first reduce the frequency of requests and increase the pool of IPs without changing the type.
- If captcha appears for new accounts — check the correspondence between the geolocation of the IP and the settings of the anti-detect browser (timezone, language, screen resolution).
- Test changes on 2-3 accounts or requests before mass implementation.
Conclusion and recommendations
The error code is not just a technical detail, but a direct hint from the site about the reason for the block. 403 indicates that the type of IP or its combination with the browser fingerprint is not suitable. 429 indicates the speed of requests and requires rotation and delays, not a change in the category of proxies. Captcha signals behavioral and reputational problems, which are solved by a combination of proxies and correct simulation of user actions.
If you regularly see 403 when working with social networks or advertising accounts, start by switching to residential proxies — they provide IPs of real users and minimal risk of immediate blocking. For farming accounts and tasks where maximum similarity to regular mobile traffic is important, mobile proxies are more suitable. And if your main problem is 429 during mass data collection from marketplaces, where speed and volume are important, it is more cost-effective to use data center proxies with properly configured rotation. Choosing the right type of proxy for a specific error code saves budget and frees you from the endless cycle of "changed proxy — banned again."