perf(integrate): run the embedded agent on Haiku 4.5, raise per-run budget to $10 - #10
Merged
Merged
Conversation
…aiku), budget $10 The embedded integration agent ran on one model with a $5 per-run cap; a full Customer Portal integration exhausted the Seam-hosted inference budget mid-run. Pick the model by the chosen mode instead: the customer-portal path is a small, high-value surface (embedded portal, reservation deep-links) and runs on Opus 4.8; the broader full-API path runs on cheap Haiku 4.5. Raise maxBudgetUsd to $10. Pairs with seamapi/seam-connect#17059 (proxy monthly cap -> $10; meter is a per-model price map so each path is priced by the model that actually ran).
itelo
force-pushed
the
itelo/cheap-model-budget-10
branch
from
August 11, 2026 19:11
20850fa to
3d069d4
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
A full Customer Portal integration hit the Seam-hosted inference budget mid-run (
429 inference_budget_exhausted,limit_usd: 5).Changes to the embedded integration agent in
src/lib/steps/integrate.ts:modelForMode): the customer-portal path — a small, high-value surface (embedded portal + reservation deep-links) — runs onclaude-opus-4-8; the broader full-API path runs on cheapclaude-haiku-4-5. Defaults to Haiku when the mode is unknown.modeis already passed intorunIntegration.maxBudgetUsd5 → 10 — headroom for a multi-flow integration.Pairs with seamapi/seam-connect#17059, which raises the proxy's per-workspace monthly cap to
$10and makes the cost meter a per-model price map so each path is priced by the model that actually ran.Testing
tsctypecheck cleanSecurity & Compliance