#does warp has the "us-central1" region hard-coded inside the app?

4 messages · Page 1 of 1 (latest)

muted walrus Dec 11, 2025, 10:27 PM

im getting 429 error back to back with my gemini Api. after searching a great deal and talking to gemini it self (+changing the system environment variables for both vertex ai location and google cloud location) the only thing that is left is that Warp is ignoring my computer's settings and has the "us-central1" region hard-coded inside the app.
its calling vertex ai when im not using that one and im using the ai studio key ...
i even created a new one (new project + new key) and used that one and i also got 429 after.
so if that is true and the region is hard coded you guys should change it ,,,
PS : this started happening after your Api key prolem fix. which was from your back end and got resolved 2 days ago (im not saying they are related but timing is on point)

next lichenBOT Dec 11, 2025, 10:27 PM

Hi @muted walrus. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

I’m sorry, but the knowledge sources don’t contain any explicit information about which region Warp uses for Gemini / Vertex AI, or whether us-central1 is hard‑coded anywhere. They also don’t describe how Warp chooses between Vertex vs AI Studio endpoints, or how (or if) it respects local environment variables for region selection.

What we do know from the docs and recent issues:

Warp supports BYOK for Google models and routes requests “directly to the model provider you’ve configured,” but there’s no mention of region configuration or overrides. [BYOK]
There was a recent backend issue specifically affecting custom Gemini keys that Warp fixed on their side, without requiring an app update. [Gemini BYOK issue]
Other users have reported Gemini / Vertex‑related errors and 4xx responses that were ultimately traced back to Warp’s backend rather than their own configuration. [Gemini BYOK issue; Vertex AI error]

None of these sources confirm:

Whether Warp is hard‑coding us-central1.
Whether Warp is ignoring GOOGLE_CLOUD_LOCATION / Vertex region env vars.
Whether the current 429s you’re seeing are a new regression tied to the recent BYOK backend fix.

Given that:

This behavior appears only in Warp and not in other tools (similar to other Gemini/Vertex reports). [Gemini BYOK issue]
There was a very recent Gemini/BYOK backend regression that Warp fixed server‑side. [Gemini BYOK issue]

the most likely path forward is:

Capture a fresh debugging ID for a failing Gemini request (from the AI FAQ instructions) and
Open or update a GitHub issue with that ID and a note that you’re seeing repeated 429s on Gemini BYOK after the recent backend fix.

That’s exactly how the previous Gemini BYOK bug was identified and resolved. [Gemini BYOK issue]

Because the docs and issues don’t expose Warp’s internal region logic, I can’t confirm or deny that `us-central1` is hard‑coded or that your env vars are being ignored. The Warp team will need to confirm this directly once they have your logs / debugging ID.

#does warp has the "us-central1" region hard-coded inside the app?

Because the docs and issues don’t expose Warp’s internal region logic, I can’t confirm or deny that us-central1 is hard‑coded or that your env vars are being ignored. The Warp team will need to confirm this directly once they have your logs / debugging ID.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

Because the docs and issues don’t expose Warp’s internal region logic, I can’t confirm or deny that `us-central1` is hard‑coded or that your env vars are being ignored. The Warp team will need to confirm this directly once they have your logs / debugging ID.