After using 15,000 credits, here is my honest opinion of Teable.

Overall, the concept of a database with AI is interesting. It is nice that it can create dashboards from a table you build or upload from a CSV. For small datasets (less than 100 rows), it seems less error prone.

Now the negatives note I used their best model only in my tests (Opus 4.6):

1. It cannot handle large datasets. I added a 30,000-row dataset, and it was unable to follow my instructions to clean or reorganize any of the data. It kept stalling and stopping without doing anything, even though it burned through 1,000 credits every time I tried to click the retry button.
2. The platform is not stable. Multiple times, the AI chat stalled and did not continue following my instructions, even when I tried different use cases. Again, it burned through credits without solving my problem.
3. The credit burn is real. If the mistake is on the AI, why does it consume my credits even when it couldn’t solve my problem? Why don’t we get credits refunded when the AI just stops or stalls?
4. It is very slow. Waiting for the AI to respond or analyze can take over 13 minutes sometimes, and when it stalls and does nothing, this is the most frustrating part as you restart or re ask the same instructions and waste more time.
5. I am pretty good at prompting, but even when the AI follows instructions without stalling, it can hallucinate and sometimes add incorrect or irrelevant information you did not ask for, so you have to double-check every time.
6. I appreciate the upfront credits I got, but I have used them all, and most cases I tried were not successful due to the AI stalling and not following instructions.
7. The context window for the AI is set to 200,000 and you cannot change it. In my opinion, this is too small and could be why the AI stalls with big datasets or certain use cases, since it doesn’t have enough space to cover everything.

I am seriously considering refunding this tool. I am happy some other users have enjoyed it, but for me it has been the complete opposite. Maybe my use cases and datasets are too large to handle, I don’t know, but Teable just did not solve most of my cases.

-------------------------------------------------
Update:
@Gary I bought the tool yesterday so Ive used it yesterday and today and I believe this period is post your updates on the models and agents. I did send a support ticket yesterday via Teable but no response from your team, I did check no email in my spam I just got an automated response below but no actual response yet on my issue:

"We will notify you of any status updates. If you have more details about this issue, please reply to this email. To keep this thread focused, kindly avoid unrelated topics—thank you.

Best regards,

Teable Support Team"

-------------------------------------------------
Update 2: (Email from them asking me to remove my review)

From: Leo

Hi,

Thanks for sharing these details.

It also looks like our previous follow-up email may not have reached you properly. We had sent it to gather a bit more detail so we could better assess the issue, and I’ll attach that message here as a .eml file as well for reference.

In your case, this appears to be related to an issue we’ve identified where the Agent may get interrupted when handling more complex tasks. The Agent is still in beta, and improving this is now one of our top priorities.

To help us better understand the impact on your side, could you share roughly how much of your prompts the AI was actually completing in practice? For example, was it usually closer to 20%, 50%, or 80% of what you asked? Any additional feedback on where it became unreliable would also be very helpful for us.

Based on the information we have so far, we believe your 2-star review may be related to this issue. If you’re open to it, we’d appreciate you considering removing or updating the review for now while we work on the fix. Once this is improved, we hope you’ll reevaluate your experience with Teable based on the progress we’ve made.

Best regards,

Leo
Product Consultant | Teable
---------------
My reply:

I would say the AI completed 30-50% of the task and got stuck or stopped after that.

In the review, I shared my honest experience. It is not my fault the AI did not work for me, yet I still lost 80% of my credits. So why would I remove my review or give you a 5 star rating? I spent $500 on your tool, which is a lot of money, but you ask me to remove my review? Sorry, that is not fair, Leo or fair on my experience and money spent.