There are several issues with the API documentation and pagination for the training text endpoint that need to be addressed:
1. Documentation Doesn't Match Actual Response
The endpoint GET /api/v1/chatbots/{chatbot_id}/training-texts returns a different response than what's documented at:
The actual response doesn't include the "content" field, making it impossible to compare training text content directly. I had to work around this by comparing the event modification date on the website with the text snippet "updated_at" timestamp.
2. Pagination Not Documented
The API only returns the first 25 text snippets per request and requires pagination, but this is nowhere to be found in the documentation. This should be clearly highlighted.
3. Inconsistent Pagination Limits
The API returns all 434 tags at once, but only 25 text snippets per request. For 434 training texts, requesting them 25 at a time might put even more strain on your server than allowing to return all at once—especially since "content" is not included, making it a rather lightweight JSON anyway.
Suggestion: Even returning 100 snippets per request would decrease processing time significantly on both your end and ours.