Release v3.7.3 by diegosouzapw · Pull Request #1724 · diegosouzapw/OmniRoute

diegosouzapw · 2026-04-28T12:04:45Z

[3.7.3] — 2026-04-28

🐛 Bug Fixes

fix(claude): strip existing billing headers from system array before injecting to prevent Anthropic prompt cache misses — stacked x-anthropic-billing-header blocks invalidated prefix matching, causing ~100% cache_create instead of cache_read ([BUG] OmniRoute stacks and modifies x-anthropic-billing-header in system prompt, breaking Anthropic Prompt Caching #1712)
fix(claude): strip output_config.format for non-Anthropic Claude-compatible providers during passthrough — third-party Claude endpoints (MiniMax, DeepSeek via aggregators) reject structured output fields with 400 errors (fix: strip output_config.format for non-Anthropic Claude-format endpoints #1719)
fix(combo): set terminal error state on response quality validation failure — prevents misleading ALL_ACCOUNTS_INACTIVE 503 when the real issue is response quality validation (fix(combo): false ALL_ACCOUNTS_INACTIVE when single-target response fails quality validation #1707, fix(combo): avoid false ALL_ACCOUNTS_INACTIVE on quality failures #1710)
fix(combo): treat combo fallback as target-level orchestration — all non-ok responses (including generic 400s) now fall through to the next target instead of being terminal; removes complex bad-request allowlist regex (fix(combo): fall back across targets on all 400 responses #1713)
fix(codex): restore namespace MCP tools and hosted-tool whitelist — regression from Fix Codex Responses WebSocket memory retention and weekly limit handling #1581 that silently dropped all MCP tool groups and Responses-API hosted tools (fix(codex): restore namespace MCP tools + hosted-tool whitelist (regression from #1581) #1715)
fix(codex): add neutral instructions for bare chat requests — Codex Responses backend rejects requests without instructions, making Codex unusable for normal chat ([Urgent] fix: add neutral instructions for bare chat in Codex provider #1709)
fix(proxy): wrap proxy assignment queries in try-catch for missing proxy_assignments table — Electron installs where migration 004 hasn't run no longer crash with no such table error ([BUG] no such table: proxy_assignments #1706)
fix(migration): improve Windows file URL path resolution in migration runner — adds direct URL path extraction and process.cwd() fallback for CI-built bundles with leaked build-time paths ([BUG] Internal Server Error 3.7.2 #1704)
fix(ui): fix light mode active request payload modal — add missing --color-card theme token, use opaque bg-surface instead of translucent bg-card/70, add backdrop blur (Fix light mode active request payload modal #1714)

🔄 Updates

chore(image-models): refresh image generation model registry — replace stale FLUX aliases with FLUX Kontext / FLUX.2 mappings, remove deprecated FLUX Redux/Depth/Canny variants (Update Image Generation Models list #1722)

Tests

All tests pass (3803 passing)

⚠️ After merging: run Phase 2 steps to tag, publish, and deploy.

# Conflicts: # tests/integration/chatcore-compression-integration.test.ts # tests/integration/resilience-http-e2e.test.ts

gemini-code-assist

Code Review

This pull request updates the application to version 3.7.3 and performs a significant refresh of the image generation model registry, updating model aliases and provider configurations for OpenAI, Black Forest Labs, and Stability AI. It also includes several bug fixes related to combo fallback orchestration, Codex tool restoration, and UI modal styling. Review feedback identifies typos in comments, inconsistent casing for model names, and the use of confusing placeholder names for inactive providers. Additionally, a discrepancy was found in the test suite where an unsupported image size was used for the OpenAI provider.

gemini-code-assist · 2026-04-28T12:06:51Z

    supportedSizes: ["1024x1024"],
  },

+  //Curruntly no models serving


Typo in comment: 'Curruntly' should be 'Currently'.

Suggested change

//Curruntly no models serving

//Currently no models serving

gemini-code-assist · 2026-04-28T12:06:51Z

-      { id: "black-forest-labs/flux-dev", name: "FLUX.1 Dev" },
-    ],
-    supportedSizes: ["1024x1024", "512x512"],
+    models: [{ id: "black-forest-labs/flux-schnell", name: "No Model yet" }],


Using 'No Model yet' as a model name is confusing for users. If the provider is temporarily unavailable, it is better to use an empty array or a dedicated status flag to handle this in the UI, rather than providing a dummy model that might still be selectable.

gemini-code-assist · 2026-04-28T12:06:51Z

+    models: [{ id: "SDXL1.0-base", name: "No Model yet" }],
+    supportedSizes: ["1024x1024"],
  },
+  //Curruntly no models serving


Typo in comment: 'Curruntly' should be 'Currently'.

Suggested change

//Curruntly no models serving

//Currently no models serving

gemini-code-assist · 2026-04-28T12:06:52Z

+      { id: "flux-2-klein-9b", name: "flux 2 Klein 9B" },
+      { id: "flux-2-klein-4b", name: "flux 2 Klein 4B" },


Inconsistent casing for the model names. Other models in this list use 'FLUX.2' (e.g., line 401). Please ensure consistent naming across the registry.

Suggested change

{ id: "flux-2-klein-9b", name: "flux 2 Klein 9B" },

{ id: "flux-2-klein-4b", name: "flux 2 Klein 4B" },

{ id: "flux-2-klein-9b", name: "FLUX.2 Klein 9B" },

{ id: "flux-2-klein-4b", name: "FLUX.2 Klein 4B" },

gemini-code-assist · 2026-04-28T12:06:52Z

        prompt: "city skyline",
        n: 2,
-        size: "1024x1792",
+        size: "1024x1536",


The size '1024x1536' is not in the supported sizes list for the 'openai' provider in imageRegistry.ts (which only lists '1024x1024', '1024x1792', '1792x1024'). This might cause validation failures in the test suite.

size: "1024x1792",

gemini-code-assist · 2026-04-28T12:06:52Z

      prompt: "city skyline",
      n: 2,
-      size: "1024x1792",
+      size: "1024x1536",


The size '1024x1536' is not supported by the 'openai' provider. Use a supported size like '1024x1792' to match the expected upstream body and the registry configuration.

size: "1024x1792",

github-actions · 2026-04-28T12:09:55Z

CI Coverage Report

Coverage job: success
PR test policy: success

Coverage artifact was not available for this run.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d2da66c339

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-04-28T12:12:31Z

  "pollinations/kontext": {
-    provider: "pollinations",
-    model: "flux-kontext",
-    name: "FLUX Kontext",
+    provider: "black-forest-labs",
+    model: "flux-kontext-pro",


Keep pollinations/kontext routed to Pollinations

This alias now maps an explicitly Pollinations-scoped model ID to black-forest-labs, so parseImageModel("pollinations/kontext") resolves to a different provider than the caller requested. In the image route, credentials are loaded by parsed provider, so users who only configured Pollinations now get No credentials for image provider: black-forest-labs for a model string that previously used Pollinations credentials. This is a backward-compatibility break for existing pollinations/kontext clients.

Useful? React with 👍 / 👎.

chatgpt-codex-connector · 2026-04-28T12:12:31Z

-        inputModalities: ["text", "image"],
-        description: "Higher quality Kontext editing with optional source image",
-      },
+      { id: "gpt-image-2", name: "GPT Image 2" },


Avoid introducing ambiguous bare model ID gpt-image-2

Adding gpt-image-2 to Pollinations duplicates an existing OpenAI model ID, and the no-prefix fallback in parseImageModel picks the first provider match by registry order. That means requests using model: "gpt-image-2" will always resolve to OpenAI, which can misroute traffic and produce credential errors in setups that intend to use Pollinations without prefixing. Keep bare IDs provider-unique (or require explicit provider/model).

Useful? React with 👍 / 👎.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c5213d4415

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-04-28T13:07:20Z

+    model: "flux-2-max",
+    name: "FLUX.2 Max",
    listInCatalog: true,
    inputModalities: ["text", "image"],


Drop image-required modality from FLUX.2 catalog aliases

Marking flux-2-max (and the sibling flux-2-pro / flux-2-flex aliases added in this change) as inputModalities: ["text", "image"] makes the route layer require both prompt and image input (requiresImageInput check in src/app/api/v1/images/generations/route.ts), so prompt-only requests to these new catalog aliases now fail with Image input is required... even though the Black Forest Labs handler accepts text-only generation for these models. This turns the primary text-to-image path for the new aliases into a 400 unless callers provide an image.

Useful? React with 👍 / 👎.

sonarqubecloud · 2026-04-28T13:07:57Z

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

* Update image-model list * docs: update CHANGELOG.md with merged PR entries for v3.7.3 * chore(release): bump to v3.7.3 — changelog, docs, version sync * chore(workflow): update generate-release with Phase 2 validation step --------- Co-authored-by: backryun <bakryun0718@proton.me> Co-authored-by: diegosouzapw <diegosouzapw@users.noreply.github.com>

backryun and others added 4 commits April 28, 2026 20:31

Update image-model list

76c04a8

Merge branch 'release/v3.7.3' into backryun/main

b0979d7

# Conflicts: # tests/integration/chatcore-compression-integration.test.ts # tests/integration/resilience-http-e2e.test.ts

docs: update CHANGELOG.md with merged PR entries for v3.7.3

11a0bcf

chore(release): bump to v3.7.3 — changelog, docs, version sync

d2da66c

gemini-code-assist Bot reviewed Apr 28, 2026

View reviewed changes

chatgpt-codex-connector Bot reviewed Apr 28, 2026

View reviewed changes

chore(workflow): update generate-release with Phase 2 validation step

c5213d4

diegosouzapw merged commit ae8a7d2 into main Apr 28, 2026
66 checks passed

chatgpt-codex-connector Bot reviewed Apr 28, 2026

View reviewed changes

		{ id: "flux-2-klein-9b", name: "flux 2 Klein 9B" },
		{ id: "flux-2-klein-4b", name: "flux 2 Klein 4B" },

Uh oh!

Conversation

diegosouzapw commented Apr 28, 2026

[3.7.3] — 2026-04-28

🐛 Bug Fixes

🔄 Updates

Tests

⚠️ After merging: run Phase 2 steps to tag, publish, and deploy.

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

gemini-code-assist Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented Apr 28, 2026

CI Coverage Report

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

sonarqubecloud Bot commented Apr 28, 2026

Quality Gate passed

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants