Skip to content

[LLMO] Redact inline base64 images from input #305

@Radu-Raicea

Description

@Radu-Raicea

We need to replicate these changes in the Python SDK for LLMO: PostHog/posthog-js#2217

The original issue: PostHog/posthog-js#2134

Allowing inline base64 encoded images to be sent as part of the PostHog event is not safe due to the maximum size of the event.

A better approach is to first upload the image via some sort of Files API, and then use its URL. For those that still want to use inline base64 encoded images, they need to be redacted from the PostHog event.

Metadata

Metadata

Labels

bugSomething isn't working

Type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions