The AI assistant in your TYPO3 backend.
Say in plain words what you need. ChEddi gets it done right where you are working.
Meet ChEddi
ChEddi sits as a chat on every page of your TYPO3 backend. You describe what you need, it takes care of the rest: finding pages, creating content, translating, filling in metadata.
It works with your permissions, asks before every change and puts everything into a draft first. You stay in control, it takes the clicking off your hands.
What changes day to day
Not just faster. Different.
Four points where you decide
Between your question and the published page there are four checkpoints. ChEddi skips none of them.
With your permissions
ChEddi acts as the logged-in editor. It only sees what they may see and only changes what they may change. Anyone without access to a page does not get it through the chat either.
Reading runs, writing asks
Searching and looking things up happens without asking. As soon as something is to be saved, ChEddi stops and shows you what it intends to do.
A preview before every save
Old value, new value, record by record. Only your click triggers the save. For hard-to-reverse actions you need two.
Draft first, live second
Approved changes land in the workspace. Visitors see none of it until you publish. Or discard.
Your content, your rules
The GDPR mode of the AI Suite applies in the chat as well.
Ready to go in 3 steps
No client, no token, no connection to set up. ChEddi runs in the backend you already have.
-
1
Install
Pull ChEddi into your TYPO3 via Composer.
-
2
Enable
Enable the backend group and assign the models.
-
3
Get going
Open the chat and give it your first task.
Working with the AutoDudes exceeded our expectations from the very start. The practical benefit was noticeable immediately. What used to take hours we now solve in minutes, translations above all. What impressed us most was the support: fast, competent and always at eye level. You can tell that real experts are at work here, people who do not just deliver technology but genuinely think along. The collaboration feels less like a classic service relationship and more like a reliable team growing alongside us.
Frequently asked questions about ChEddi
Do I need a subscription of its own?
No. ChEddi is part of the AI Suite and requires it. If you already use the AI Suite, all that is added is one extension.
Does ChEddi see my screen?
No. ChEddi sees neither your screen nor the view you have open. If you have a record open, it is told that record's table and number, nothing more.
What happens if ChEddi gets something wrong?
Nothing is saved without confirmation, and confirmed changes land in the draft workspace by default. From there each one can be discarded individually.
Which data leaves your installation?
Your message, the conversation so far and the text from attachments go to the AI Suite Server, which runs the model. The chat never talks to an AI provider directly.
With GDPR mode active, the model selection is limited to models marked as GDPR-compliant, and web research runs through a search provider hosted in the EU.
Is web research available straight away?
No, and that is deliberate. It ships switched off and has to be enabled per backend group. That way you decide yourself who may send search queries out.
With GDPR mode active, the search runs through a provider hosted in the EU. If that provider does not serve the language of your query, the search is skipped and the chat states the reason. Reading an address you name yourself is not affected.
Does every request cost credits?
Yes, every request consumes credits from your AI Suite allowance. Billing is by tokens, and the price is shown with each model in the selection. Your balance is visible at the top of the chat at all times.
Can I control who may use ChEddi?
Yes, per backend group. Without approval the chat does not even appear. On top of that you decide which AI models a group may use.
Is ChEddi finished yet?
ChEddi is still in beta. That means: usable in production, but the feature set is still growing. We recommend pinning a fixed version and reviewing the changelog before updates.