Vapi is the Voice AI platform for developers.

Vapi AI

I recently migrated to Trieve for my knowledge base provider. I'm using it through VAPI API of creating knowledge base which is assigned by default. I wonder, what's the chunk limit for Trieve cause some files are getting indexed and some don't 🤷

Heyarmandas Thank you for creating the ticket! A member of our team will reach out to you shortly.

<@880748155241652234>
I'm glad to hear you're using Trieve through the VAPI API! The chunk limit for Trieve can vary based on the specific implementation and the plan you are subscribed to. Typically, there might be a limit on the size of individual documents or the number of chunks that can be indexed in a single operation.

To get the most accurate information regarding the chunk limit for Trieve, I would recommend checking their official documentation or reaching out to their support team. They will be able to provide you with the specifics based on your current subscription and usage.

If you have any other questions or need further assistance with VAPI, feel free to ask!

You can continue this conversation by mentioning me in the message.

It's 1000 chunks on the free tier, I wonder if it's the same in VAPI. In any case, is it possible to increase the chunk limit on VAPI?

<@880748155241652234>  Trieve’s default “chunk size” is 20 splits (sentences or segments) per chunk, and there isn’t a hard max on the total number of chunks. When you send a file to Vapi-Trieve, it splits the text into chunks of around 20 segments by default, and then rebalances them if needed. You can tweak that with `targetSplitsPerChunk`, `splitDelimiters`, and `rebalanceChunks`. 

If you notice certain files not getting indexed, it’s usually not because there’s a strict chunk limit. It could be something else like file format issues, file size limits, or others.

Could you share the file ID so I can check our logs, why a file might be skipped.

thanks for coming back. File id: 4846e576-6660-4135-9218-90621272bdfe

`logs`
--not-indexd
🔵 20:22:41:747 Processing `file` Creation...: Jan 23, 2025 at 1:52:42.262 am(9 days ago) 
--indexed
🔵 20:22:42:099 Querying `db` (connection#  28)...
Query: UPDATE file

Rest of logs are related to Canonical KB.

<@880748155241652234> Could you please provide the file ID associated with Trieve KB or the Trieve KB ID linked to this file?

Because for me to look into the dept, I required to get the DataSetID.

<@880748155241652234> Because of no response closing this ticket.

Hey 🙏 I’ll send the id in a couple of hours

file ID: bd250b2a-d746-45d8-99c7-005c7c0c3768