[waleedlatif1]

Summary

• Coalesce and cache historical fal prices per credential and endpoint, respect rate-limit cooldowns, and verify exact receipts belong to the requested job.
• Distinguish terminal queue failures from failed delivery. Start exact receipt tracking at terminal queue status, bound result retrieval, and retain interrupted-delivery reconciliation without estimating charges for explicit model failures. Report bounded, redacted provider error types.
• Allow remote generated images up to the existing 100 MiB buffered-transfer limit while retaining inline-image and JSON limits.

Type of Change

• Bug fix

Testing

Passed 98 focused provider-boundary tests; full root bun run test with VITEST_MAX_WORKERS=6 (3,455 app files, 39,525 app tests); lint; all 62 audits; workspace type-check; test-type ratchet; block-registry check; and generated-artifact checks. Regression cases cover concurrent pricing, credential isolation, cache expiry, Retry-After recovery, cancellation, receipt identity, rejected results, and declared/streamed image limits. Twelve guard-removal checks and reproduced cancellation, terminal queue-error, and concurrent-cooldown regressions demonstrate that the regressions fail when their corresponding protections are removed.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing (new tests pass the test-audit authoring gate)
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
docs	Skipped		Nov 5, 2026 10:22am UTC

[greptile-apps]

Confidence Score: 5/5

[High impact] The PR appears safe to merge, with no new actionable issues found.

Summary

This PR bounds Fal.ai pricing and result requests while preserving charge records when delivery fails.

• Fal image and video jobs keep charges visible when media delivery fails.
• Fal price lookups match receipts to jobs and reuse account-specific estimates.
• Remote generated images can use the existing 100 MiB transfer limit.

Both earlier concerns are addressed: delivery failures retain reconciliation metadata, and successful shared prices are cached even when their only waiting caller stops.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[Queue completes] --> B[Start exact receipt lookup]
  A --> C{Queue reports an error?}
  C -->|Yes| D[Skip result request]
  C -->|No| E[Fetch result within 30 seconds]
  E --> F{Explicit model failure?}
  F -->|Yes| D
  F -->|No| G[Keep result or delivery error]
  B --> H[Await exact receipt]
  D --> H
  G --> H
  H --> I{Receipt available?}
  I -->|Yes| J[Keep exact charge]
  I -->|No explicit failure| K[Use estimate or fallback floor]
  I -->|Explicit failure| L[Do not estimate]

Loading

_{Reviews (2) · Last reviewed commit: "fix(falai): preserve receipts across res..." · Reviewed by Greptile}

[cubic-dev-ai]

All reported issues were addressed across 10 files

_{Turn on auto-fix | Re-trigger cubic}

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

No issues found across 10 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}