docs: custom routers now support BYOK (still no BYOE)#336
Conversation
Custom model routers now apply BYOK after task-time model resolution (warp-server #12382). Update the custom routers page and related BYOK, custom inference endpoint, and team-managed keys pages: - Custom routers: BYOK key feature bullet; rewrite the credits section to explain per-request BYOK billing, the all-models-covered rule for running without Warp credits, team-managed keys, and that custom inference endpoints still can't be routing targets (no BYOE). - BYOK page: note custom routers as the exception to Auto-model BYOK behavior. - Custom inference endpoint page: clarify routers never resolve to endpoint-routed models. - Team-managed keys page: routers route through team/member first-party keys after resolution. Co-Authored-By: Oz <oz-agent@warp.dev>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
I'm starting a first review of this pull request. You can view the conversation on Warp. I completed the review and no human review was requested for this pull request. Comment Powered by Oz |
![oz-for-oss[bot]](https://avatars.githubusercontent.com/in/3450139?s=60&v=4){kind=small}
There was a problem hiding this comment.
Overview
This PR updates the BYOK, custom router, custom inference endpoint, and Enterprise team-managed key documentation to reflect that custom routers can apply provider API keys after resolving to concrete Warp-supported models while still excluding custom inference endpoints. I found no blocking correctness, documentation-structure, link-target, security, or spec-alignment issues in the attached diff.
Concerns
- None.
Verdict
Found: 0 critical, 0 important, 0 suggestions
Approve
Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).
Powered by Oz
1 participant
Summary
Custom model routers now support BYOK: the server applies your provider API keys after a router resolves a task to a concrete model (warp-server #12382, released last week). Custom inference endpoints (BYOE) still can't be routing targets. This PR updates the docs, which previously said custom routers never draw on your own keys.
Behavior verified against warp-server (
logic/ai/multi_agent/agents/primary/agent.go,logic/ai/multi_agent/request/model_config.go):Changes
custom-routers.mdx— Added a BYOK support bullet to Key features, and rewrote the credits section (now "Using your own API keys (BYOK)" + "Model availability and fallback"): per-request BYOK billing, the all-models-covered rule for running without Warp credits, team-managed keys, and that custom inference endpoints still can't be routing targets.bring-your-own-api-key.mdx— Auto Model section: noted custom routers as the exception — keys are applied after the router resolves a model.custom-inference-endpoint.mdx— Auto routing section: clarified routers never resolve to endpoint-routed models, while BYOK provider keys do apply.team-managed-keys-and-endpoints.mdx(enterprise) — Billing behavior: custom routers route through team/member first-party keys after resolution; routers can't target team endpoint models.Additional context
npm run build(347 pages), style lint on changed files (only pre-existing warnings on untouched lines), internal link check (0 broken).Co-Authored-By: Oz oz-agent@warp.dev
Conversation: https://staging.warp.dev/conversation/ad441006-8c0c-4389-8500-450f6df36673
Run: https://oz.staging.warp.dev/runs/019f6bd3-b02d-7bfb-a87e-b7c7021e0cdc
This PR was generated with Oz.