AI Toolbox 3.0 - Search, organize and export every AI chat in one place
by•
AI Toolbox is the missing layer on top of ChatGPT, Claude, Gemini and Grok: folders and subfolders, full-text search across every conversation (even all four platforms at once), a prompt library with // insert and multi-step chains, bulk export to Markdown/PDF/JSON, bookmarks, and a live context meter. One install covers all four AIs. Local-first: your chats stay in your browser. Trusted by 40,000+ users in 150+ countries, rated 4.5 stars on the Chrome Web Store.
Replies
I've been using the tool with Claude for about a month. The export chat to JSON has been very helpful for reviewing outstanding issues in long-running chats. Search is sometimes flaky -- doesn't find known values in chats. Context meter is a "guess" at best. Overall highly recommend.
@ipahophead thank you for this comment, regarding all issues that you experience, could you please send us an email so we can inspect those? support@infi-dev.com
If you work inside most any AI platform there is a Personalization section go there and tell your version of their AI how you want it to responed back in you want sassy, or Hard core Engineer of this or these fields and never Hallucinate double and even sometimes Triplecheck your work before dishing it up for me to digest. It's all about how Tell it to treat you & what you Expect from it ....if you just signed up and dove in then Losing your work is envidible mine are usually named on the first topic unless I manually change mine with the rename feature that both Claude & OpenAI have them usually those 3 dots or a right clickon the title in the chat libaray bar.
Now if your's does all this and retreives full transcripts can it translate a massive Json Packet back into full conversations?
And You Don't Copy &/or Use our Data to Train your AI or allow 3rd or even 4th Parties to view it?
Because, I've already had that issue with 1 AI Provider & I'm not in the mood to make others Wealth again while, I get Screwed !!
@tallan_acalin good questions, regarding privacy, all conversation data is stored locally on your browser, we dont use it or sell it, you are the only one with the access to your conversations.
The exported JSON files can be given back to the AI whenever needed.
wow. very useful tool that enhances the use experience. Curious what are the updates compared with the previous versions.
@jamieg Thanks! You can check our changelog here - https://www.ai-toolbox.co/changelog
Really useful idea Ive often struggled to find old AI conversations across different tools The search and folder system makes this much easier to manage and reuse past work
@shivam_kushwaha16 Yes, the cross platform search is an amazing feature, one of our best!
@adi_leviim Yeah I can see why it stands out being able to search across all the AI tools from one place is genuinely useful
@shivam_kushwaha16 glad you find it useful :)
Great @adi_leviim
I've been juggling ChatGPT and Claude for months and losing old chats was driving me nuts. The cross-platform search alone would fix that.
Trying this today 🙌
@boyuan_deng1 Glad to hear, tell me how it worked for you :)
Congrats on the launch! It's always great seeing how people innovate around existing solutions and create new user value on top 👏 btw, I'm curious to hear if after seeing people actually use AI Toolbox, was there any user behavior or use case that genuinely surprised you? Maybe something you didn’t originally build the product for, but users ended up doing anyway? these kind of unexpected behaviors often reveal interesting product opportunities and worth keeping an eye on.
@hasanaboshally Thank you for the nice words.
The things that most surprised us is that we get on a daily basis a lot of feature requests and each one is totally different from the other.
It just shows us how much room we still have to make this product scale and a lot larger.
the local-first caching makes sense for privacy, but curious how brittle it is - ChatGPT/Claude/Gemini all change their DOM structure pretty often. when one of them ships a UI update, does the extension quietly break for that platform until you push a patch, or is the scraping resilient to that by design?
@galdayan Good question, and many users have asked the same question hundreds of time - the only features that might break are items that we add inside a chat menu, and we immediately fix those and a fixed version is live - it haven't happened for a long time and everything is 100% stable now.
All the other features don't get affected from their UI changes so the product is good to be used :)
@adi_leviim good to know, and honestly makes sense - the chat UI itself is the volatile surface, everything downstream of that (search, folders, export) is your own infrastructure so it doesn't care what ChatGPT's sidebar looks like this week.
Local-first search across four platforms means that index lives in IndexedDB, and Chrome evicts that under storage pressure without telling anyone. Two years of history plus a folder tree is exactly the thing that quietly vanishes and then looks like a bug in your extension. Are you calling navigator.storage.persist() and surfacing the quota anywhere, or is the context meter the only storage number in the UI? I'd put that in front of the folders, since losing the organisation hurts more than losing a chat.
@asadmalik901 Chrome can remove storage if there is not enough space to clear things up so your browser will not get stucked.
However, the conversations, even if you have hundreds of thousands, do not weigh too much to make Chrome to clear your IndexedDB.
The context meter is calculated by your LLM plan and the conversation content
@adi_leviim You skipped the bit I was asking about. persist() isn't about how much space you're using, it's about which bucket you're in, and without it Chrome can evict the origin under pressure no matter how light the data is. It's one call that returns true or false, so it's worth knowing which one you get. The folder tree is the part nobody can rebuild from the source site, and that's what I'd protect.
@asadmalik901 @asadmalik901 persist() is not called today, the folder tree does not live in IndexedDB, it is stored server-side, so eviction cannot lose your organisation. What lives locally is the search index, which is a cache and rebuilds from the site on the next sync. The worst case today is a re-index, not lost folders. Adding persist() still, because a re-index on a two-year history is not free.