I’ve been working with Google Cloud’s Vertex AI platform and successfully implemented custom labels for my language model requests through the generateContent API. This works great for the standard Vertex AI models.
Now I’m trying to do something similar but with Anthropic’s models instead. I’m using their official Vertex SDK package (@anthropic-ai/vertex-sdk from npm) to make the API calls.
The thing is, I can’t figure out how to attach custom metadata labels to these Anthropic model requests like I can with the regular Vertex AI calls. Has anyone managed to get this working? I need to track and organize my API usage with labels, but the Anthropic SDK documentation doesn’t seem to cover this specific use case on the Vertex AI platform.
Any guidance on whether this is even supported and how to implement it would be really helpful. Thanks!
yea, that’s a known issue with the anthro model wrapper. i’ve been intercepting requests b4 they hit the sdk and manually addin custom headers. it’s kinda hacky but works for basic labeling. u could skip the anthro wrapper and go for the vertex ai rest api directly!
Hit this same issue a few months ago when migrating workloads to Claude on Vertex AI. The Anthropic SDK wraps Vertex calls but cuts out platform features like custom labels. I ditched the Anthropic SDK for requests needing labels and went straight to Google Cloud client libraries instead. Had to format the request payload manually to match what Anthropic models want. More boilerplate code, but you get full Vertex AI features including custom labels. Worth the extra work since we needed proper cost attribution.
Currently, the Anthropic Vertex SDK does not facilitate the addition of custom metadata labels in the same way that standard Vertex AI models do. I encountered a similar issue previously while transitioning to Anthropic’s SDK. Since the @anthropic-ai/vertex-sdk package tends to obscure the underlying Vertex AI mechanisms, accessing features like custom labeling becomes problematic. As a workaround, I implemented a tracking system that logs the necessary metadata alongside request IDs to a separate database. This approach allowed for effective cost tracking and usage analysis, although it is not as seamless as integrating labels directly. Alternatively, you might consider making direct REST API calls to the Vertex AI endpoint for enhanced control over your request attributes, although I have not personally tested that route.
Hit this same issue last week on a project where we needed detailed billing breakdowns for different client requests.
I ditched the Anthropic SDK completely and went with the Google Cloud Vertex AI client instead, building the payloads manually. You lose some convenience but get full control over request metadata.
I built a wrapper function that takes your prompt and labels, then constructs the raw Vertex AI request with proper Anthropic formatting. Just set your custom labels in the request headers before hitting the Vertex endpoint.
Took about an hour to get working, but now all our Claude requests have proper labels for cost tracking. Way cleaner than database logging since everything stays in the Google Cloud ecosystem.
If you’re handling high volume requests, this approach also gives you better retry logic and error handling.
Setup’s pretty straightforward once you get how Vertex AI expects the Anthropic model requests formatted.
