Claude blocks Russian IPs: how not to lose your AI project

Claude has blocked Russian IP addresses en masse. Several hundred users lost their accounts entirely, along with months of development work: AI agents, analytics, and automations. VPNs didn't help. Here's why it happened and how to protect your project.

AI-processed from Habr AI; edited by Hamidun News
Claude blocks Russian IPs: how not to lose your AI project
Source: Habr AI. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceHabr AI ↗
Editor-in-Chief
Key metric #1Finished AI agents that wrote code and processed data
Key metric #2Custom system prompts and fine-tuned configurations
Key metric #3Conversation history and context from long-running projects

Several hundred Russian users have lost access to their Claude accounts. Even with VPNs, they couldn't get back in, because the server could see where the request was actually coming from. Months of work with AI agents, data, and finished automations are all gone.

What happened

Anthropic blocked access from Russian IP addresses. It happened without warning: users simply tried to log in and were refused. Even when they connected through a VPN, Claude saw the real IP through browser or traffic metadata, and the block stayed in place.

The exact reason is unclear. It may be an Anthropic decision driven by sanctions, a technical error, or a response to heavy traffic from Russia. The company has not issued an official statement.

What users lost

  • Finished AI agents that wrote code and processed data
  • Custom system prompts and fine-tuned configurations
  • Conversation history and context from long-running projects
  • Analytics data and experiment results
  • Access to billing accounts and usage history

Recovering data through Anthropic isn't working: support either doesn't respond or sends a stock reply saying the service is unavailable in the region.

How to protect yourself

"I no longer rely on a single platform," writes one of the affected users.

Here's what to do right now:

  • Back up your prompts and configs — keep important system prompts and agent instructions in your own documents, git repositories, or cloud storage
  • Export your history — download your Claude conversations as JSON if you can, or manually copy out key dialogues
  • Use local models for critical tasks — Llama 3.1, Mistral, or other open-source alternatives you can run on your own server
  • Diversify your tools — work with several services in parallel (Claude, GPT, Gemini, DeepSeek) and don't make them depend on each other
  • Set up your own API gateway — if you pay for API access, put your own layer between your application and the service so you can switch quickly

What it means

The Claude block in Russia shows how fragile cloud AI tools are when you rely on them completely. For freelancers and startups, it's a signal to build a more resilient tech stack: combine cloud services with local models, document and archive your work, and don't keep all your eggs in one basket.

This isn't the end of the Russian AI community, but it is a serious reason to rethink how projects are architected.

Story

Key Questions Answered

What happened?

Anthropic blocked access from Russian IP addresses. It happened without warning: users simply tried to log in and were refused. Even when they connected through a VPN, Claude saw the real IP through browser or traffic metadata, and the block stayed in place. The exact reason is unclear. It may be an Anthropic decision driven by sanctions, a technical error, or a response to heavy traffic from Russia. The company has not issued an official statement.

What users lost?

Finished AI agents that wrote code and processed data Custom system prompts and fine-tuned configurations Conversation history and context from long-running projects Analytics data and experiment results Access to billing accounts and usage history Recovering data through Anthropic isn't working: support either doesn't respond or sends a stock reply saying the service is unavailable in the region.

How to protect yourself?

"I no longer rely on a single platform," writes one of the affected users. Here's what to do right now: Back up your prompts and configs — keep important system prompts and agent instructions in your own documents, git repositories, or cloud storage Export your history — download your Claude conversations as JSON if you can, or manually copy out key dialogues Use local models for critical tasks — Llama 3.1, Mistral, or other open-source alternatives you can run on your own server Diversify your tools — work with several services in parallel (Claude, GPT, Gemini, DeepSeek) a

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Hands-On AI Engineering · Hamidun Academy

Build & Ship Production AI Agents Without Boilerplate Code

Start your free 72-hour trial at Hamidun Academy (academy.hamidun.com) — interactive lessons on Claude Code & Cursor, live Tech Radar, and production agent templates.

Start 72h Free Trial →Vibe-Coding Intensive
What do you think?
Loading comments…