Hoody AI

@grigory_sl Great question, it was only asked privately on Telegram so far, we are glad to reply publicly. We do not 'scrape' LLM providers using headless instances, instead, we use their official API and our usage is legitimate and allowed by their TOS, so it's very unlikely that our API keys would be banned. However, we already rotate a serious number of API keys from all providers to anonymize users further, protect against balance draining and also cover preemptively the scenario you've described, which means that when a user interacts with an LLM, it is likely that they will be using different API keys even if it's the same chat. The more users on Hoody AI, the more anonymous every user becomes because profiling/correlation of prompts becomes incredibly harder. The servers reaching the LLM endpoints are also heavily proxified, using our own Bulletproof Network, making IPs virtually unblockable as they are rotating constantly, it is one of our main strength as Bulletproof Proxies/VPNs is a service we already offer to our users. We will update `docs.hoody.com` to cover most of Hoody AI technical questions in the coming days. Have a great day ahead!

@kyrylosilin This is a great question! Thank you for asking. We believe that any information is already available online, and what users ask AI models is their own responsibility. Of course, we do not condone or encourage harmful or illegal activities. However, the bigger question is: Should our prompts be available to government entities and advertisers? Absolutely not! This is precisely what we've solved here. Most AI users are not yet aware that every prompt sent to AI giants (think OpenAI, Anthropic...) is permanently recorded by Google Analytics, and Google has a history of sharing user data (100,000+ users) with law enforcement agencies. As of September 2024, virtually all AI products are leaking their prompts automatically to Google and many other third parties. Governments are already very likely profiling AI users at scale. Our core principle is privacy and anonymity of our users, we strictly design privacy products that ensure we completely lose the ability to know what users are doing on our Network, that includes ALL our products, both for ethical and legal reasons.

@lumi_nertan Thanks to Hoody Keys, authentication is secure and anonymous by nature. We are using the best encryption algorithms available to date, including slow password-hashing algorithms such as Argon2. An attacker would need to try approximately 7.96 × 10^24 (7.96 septillion) combinations to guess a Hoody Key. This is infinitely more secure than any Email account that can be stolen through many ways. You can learn more about it here: https://docs.hoody.com/en/hoody-...

@911lord Thank you, our designers always love the appreciation :) We can't guarantee that Hoody AI will stay free forever, however, we intend to keep it that way for as long as we can, of course, offering free access to OpenAI, Anthropic... costs a lot, it's all about the balance between Premium users and fair usage. In the coming days, the limits of prompts, tokens... will be made public on our Pricing page. Thank you so much for signing up!

@leugomil We also believe that providing privacy to AI users is extremely important, as virtually everyone will be using AI models in the near future. Our dev team is extremely competent and all of them have a deep passion for privacy. We always focus on providing anonymity for real-life and practical use cases, so users don't have to go to extreme lengths to enjoy privacy, which should be a fundamental right. However, the current state of the internet makes it really hard for non-power users to stay private. Thank you so much for your appreciation and for supporting Hoody!