Skip to content

🐛 fix(context-engine): downgrade image_url parts when target model lacks vision#14029

Merged
arvinxx merged 3 commits into
canaryfrom
fix/downgrade-vision-messages
Apr 22, 2026
Merged

🐛 fix(context-engine): downgrade image_url parts when target model lacks vision#14029
arvinxx merged 3 commits into
canaryfrom
fix/downgrade-vision-messages

Conversation

@arvinxx

@arvinxx arvinxx commented Apr 22, 2026

Copy link
Copy Markdown
Member

💻 Change Type

  • ✨ feat
  • 🐛 fix
  • ♻️ refactor
  • 💄 style
  • 👷 build
  • ⚡️ perf
  • ✅ test
  • 📝 docs
  • 🔨 chore

🔗 Related Issue

Fixes LOBE-7214

🔀 Description of Change

Historical messages persisted in the multimodal parts form (content is an array with image_url entries, or assistant messages with metadata.isMultimodal) bypassed the legacy imageList vision check and got forwarded verbatim to the provider. DeepSeek rejects the image_url variant outright (unknown variant 'image_url', expected 'text'), so any topic containing an image broke the moment the user switched to a non-vision model.

This PR introduces a VISION_DOWNGRADE_PLACEHOLDER and applies it uniformly across every path that can emit an image_url part:

  • User message where content is already an array of parts (the LOBE-7214 trigger).
  • Assistant multimodal content dispatched through convertMessagePartsToContentParts (both metadata.isMultimodal and reasoning.isMultimodal).
  • Assistant legacy imageList path — previously the image was silently dropped.
  • User legacy imageList path — also switched from silent drop to placeholder, so behavior is consistent everywhere.

Replacing (rather than dropping) image parts keeps the conversational signal that an image existed without including content that non-vision providers reject.

🧪 How to Test

  • Tested locally
  • Added/updated tests
  • No tests needed

All 763 context-engine tests pass (bunx vitest run in packages/context-engine). New cases cover:

  • User imageList + non-vision model (single + multi-image).
  • User content already an array with image_url parts + non-vision model (direct LOBE-7214 regression).
  • User array content preserved verbatim when vision is supported.
  • Assistant metadata.isMultimodal + non-vision model → image parts replaced by placeholder text parts.
  • Assistant legacy imageList + non-vision model → placeholder appended to content.

📝 Additional Information

Only the image_url / image part types are downgraded. video_url parts are out of scope for this issue and left to the existing isCanUseVideo path.

…cks vision

Historical messages persisted as multimodal parts (content is an array
with `image_url` entries, or assistant messages with `metadata.isMultimodal`)
bypassed the legacy `imageList` vision check and got forwarded verbatim to
the provider. DeepSeek rejects the `image_url` variant outright, so any
topic containing an image broke the moment the user switched to a
non-vision model.

Replace image parts with a textual placeholder so the conversation still
carries the signal that an image was sent, without including content
non-vision providers reject. Applies uniformly across user array content,
assistant multimodal content, and legacy `imageList` paths.

Fixes LOBE-7214.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@vercel

vercel Bot commented Apr 22, 2026

Copy link
Copy Markdown

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

Project Deployment Actions Updated (UTC)
lobehub Ready Ready Preview, Comment Apr 22, 2026 4:53am

Request Review

@sourcery-ai sourcery-ai Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We've reviewed this pull request using the Sourcery rules engine

@codecov

codecov Bot commented Apr 22, 2026

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 90.00000% with 6 lines in your changes missing coverage. Please review.
✅ Project coverage is 66.92%. Comparing base (b02b727) to head (5975217).
⚠️ Report is 4 commits behind head on canary.

Additional details and impacted files
@@            Coverage Diff             @@
##           canary   #14029      +/-   ##
==========================================
+ Coverage   66.86%   66.92%   +0.06%     
==========================================
  Files        2100     2103       +3     
  Lines      179535   179831     +296     
  Branches    21192    18576    -2616     
==========================================
+ Hits       120038   120346     +308     
+ Misses      59373    59362      -11     
+ Partials      124      123       -1     
Flag Coverage Δ
app 59.57% <ø> (+0.09%) ⬆️
database 92.27% <ø> (ø)
packages/agent-runtime 79.72% <ø> (ø)
packages/context-engine 83.24% <90.00%> (+0.06%) ⬆️
packages/conversation-flow 92.40% <ø> (ø)
packages/file-loaders 87.02% <ø> (ø)
packages/memory-user-memory 74.74% <ø> (ø)
packages/model-bank 99.86% <ø> (ø)
packages/model-runtime 84.22% <ø> (ø)
packages/prompts 69.08% <ø> (ø)
packages/python-interpreter 92.90% <ø> (ø)
packages/ssrf-safe-fetch 0.00% <ø> (ø)
packages/utils 87.95% <ø> (ø)
packages/web-crawler 88.66% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Components Coverage Δ
Store 66.56% <95.26%> (+0.18%) ⬆️
Services 51.74% <ø> (+0.03%) ⬆️
Server 66.78% <ø> (+<0.01%) ⬆️
Libs 52.57% <ø> (ø)
Utils 80.59% <ø> (ø)
🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
  • 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b3792d50ae

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".

Comment on lines +166 to +169
const needsArrayRewrite = contentIsArray && arrayImageUrlCount > 0 && !canUseVision;

// Fast path: nothing to transform — plain text content passes through.
if (!hasImages && !hasVideos && !hasFiles && !needsArrayRewrite) {

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Preserve video parts when rewriting image_url array content

When needsArrayRewrite is true for non-vision models, user content arrays are forced through the rewrite path even if they already contain other multimodal parts. That path only rebuilds text plus image placeholders, so existing video_url parts are dropped from history. This is a regression from the previous fast-path behavior (which preserved the original array) and causes mixed image+video turns to silently lose video context when switching to a non-vision model.

Useful? React with 👍 / 👎.

Two tests in the app suite asserted the silent-drop behavior the
MessageContentProcessor used to exhibit for `imageList` + vision-off:

- src/services/chat/chat.test.ts
- src/services/chat/mecha/contextEngineering.test.ts

After this PR the processor appends the downgrade placeholder instead of
silently dropping the image, so the expected content grows by one line.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@arvinxx arvinxx changed the title 🐛 fix(context-engine): downgrade image_url parts when target model lacks vision 🐛 fix(context-engine): downgrade image_url parts when target model lacks vision Apr 22, 2026
…STEM CONTEXT

The placeholder stands in for an image the user actually sent, so it
should sit adjacent to the user text rather than trailing after the
SYSTEM CONTEXT metadata block. Reorder so the payload reads:

  <user text>

  [image omitted: not supported by this model]

  <!-- SYSTEM CONTEXT ... -->

Keeps the conversational flow intact and matches the semantic position
the image occupied in the original message.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@arvinxx arvinxx merged commit 4af6fdd into canary Apr 22, 2026
35 checks passed
@arvinxx arvinxx deleted the fix/downgrade-vision-messages branch April 22, 2026 05:07
@arvinxx arvinxx mentioned this pull request Apr 27, 2026
arvinxx added a commit that referenced this pull request Apr 27, 2026
# 🚀 LobeHub v2.1.53 (20260427)

**Release Date:** April 27, 2026
**Since v2.1.52:** 194 merged PRs · 17 contributors

> Introduce Heterogeneous Agent — Claude Code and Codex run as
first-class desktop runtimes, paired with a new Agent Signal package,
sharper desktop UX, and a wave of flagship model additions.

---

## ✨ Highlights

- **Introduce Heterogeneous Agent** — Claude Code and Codex run as
first-class desktop agents: subagent rendering, partial-message
streaming, multi-turn resume, terminal error surfacing, rich tool
inspectors, and runtime polish. (#14162, #13754, #14067, #14001, #13970,
#13942)
- **Screen capture & Quick Chat tray** — New desktop screen capture
overlay (macOS permission-gated) with Quick Chat tray and upload
pipeline improvements; chat input auto-focuses on overlay mount.
(#13818, #14097, #14105)
- **Desktop topic & tab UX** — Dedicated topic popup window with
cross-window sync, Cmd+W/Cmd+T tab shortcuts, TabBar polish, recent
working directories expanded to 20, and human approval notifications.
(#13957, #13983, #13972, #14036, #14092)
- **Git workflow built-in** — One-click pull/push from the branch chip,
ahead/behind badge, and submodule/worktree repo detection. (#14041,
#13980, #13978)
- **Agent Signal package** — New `@lobechat/agent-signal` runtime for
dynamic memory feedback signals, with OTel metrics and self-iteration in
Lab. (#14157, #14170, #14159, #14169, #14187)
- **New models** — Claude Opus 4.7 with `xhigh` effort tier, GPT-5.5,
DeepSeek V4 Flash/Pro with reasoning slider, Kimi K2.6, MiMo-V2.5/Pro,
gpt-image-2, Qwen3.6 Flash/Plus, and Pixverse-c1. (#13903, #14147,
#14114, #14004, #14089, #14039, #13923)
- **New providers** — OpenCode Zen, OpenCode Go, and Azure OpenAI Router
runtime. (#13943, #14064, #13823)
- **Mobile settings overhaul** — Full settings menu and responsive
profile layout for mobile. (#14019)

---

## 🏗️ Heterogeneous Agent

- Claude Code runtime, working-directory awareness, and sidebar polish.
(#13970)
- CC subagent rendering with persistent streamed text; parallel-tool
orphan fix. (#14001, #13968, #14024)
- Per-step usage persisted to each step assistant message. (#13964)
- Per-phase workflow expand defaults; full-expand toggle with
three-level expansion. (#14171, #13906)
- Hetero-mode actions bar; tool inspector polish. (#13963, #14034,
#14030)
- Codex desktop integration with rich tool rendering and devtools
preview. (#14067, #14100)
- Codex terminal error surfacing and CLI output tracing. (#14166)
- Tighten `isCanUseVision` default and add aggregator fallback. (#14172)
- Persist `ccSessionId` in topic metadata for CC multi-turn resume.
(#13902)
- CC account card, topic filter, and integration polish. (#13955,
#13942, #13950)
- Token-level deltas streamed via `--include-partial-messages`. (#13929)

---

## 🧠 Agent Signal & Self-Iteration

- New `@lobechat/agent-signal` package with dynamic feedback signals.
(#14157)
- AgentSignalRuntime wired through agent-tracing and observability-otel
metrics. (#14170, #14159)
- Self-iteration feature flag added to Lab; front-side flag check.
(#14169, #14186)
- Signal policy for receiving memory feedback dynamically. (#14187)

---

## 💬 Conversation

- Queue follow-up sends during running CC turns. (#14179)
- Persist per-topic chat scroll position; pin user message + fold long
messages. (#14191, #14056)
- Inline resend when editing last user message. (#14080)
- Disable first-block markdown streaming to prevent flicker. (#14193,
#13904)
- Prevent Markdown stream replay when vlist remounts streaming items.
(#14086)
- Stop repinning after manual scroll; unify scroll-to-user + spacer
hooks. (#14099, #14132)

---

## 📱 Platforms & Integrations

### Desktop / Electron

- Screen capture overlay, Quick Chat tray, and upload pipeline
improvements. (#13818)
- macOS permission gate for screen capture; auto-focus chat panel input.
(#14097, #14105)
- Dedicated topic popup window with cross-window sync. (#13957)
- TabBar polish: `+` button for new topic, dark theme blend, close icon
by default. (#13972, #14203, #13973)
- Recent working directories expanded from 5 to 20; submodule/worktree
repo detection. (#14036, #13978)
- Cmd+W / Cmd+T tab shortcuts and global shortcut consolidation.
(#13983, #13880)
- Linux icon configuration; human approval desktop notifications.
(#14042, #14092)

### Git Workflow

- One-click pull/push from branch chip; ahead/behind badge with
refactored GitCtr. (#14041, #13980)

### Mobile

- Full settings menu and responsive profile layout. (#14019)
- Agent route added to mobile router; mobile agent topic route
registered. (#14103, #14158)
- Session list skeleton row layout corrected. (#14040)

### Bot / Messaging

- DM strategy support; bot emoji and markdown render optimization.
(#14201, #14091, #14140)
- Slack webhook fix; bot platform setup guide reference. (#14052,
#14121)

---

## 🤖 Models & Providers

### New models

- **Claude Opus 4.7** with `xhigh` effort tier; strip temperature/top_p.
(#13903, #13909)
- **GPT-5.5**. (#14147)
- **DeepSeek V4** Flash/Pro cards with reasoning slider; cache-hit and
Pro discount pricing. (#14114, #14209, #14196, #14131)
- **Kimi K2.6** model with LobeHub-hosted card. (#14004, #14006)
- **MiMo-V2.5 / V2.5-Pro**. (#14089)
- **gpt-image-2**, **Qwen3.6 Flash/Plus**, **Pixverse-c1**. (#14039,
#13923)

### New providers

- **OpenCode Zen** and **OpenCode Go** with env-var support. (#13943,
#14064)
- **Azure OpenAI Router** runtime support. (#13823)
- Model alias mapping for image and video runtimes. (#13896)
- Seedance video models migrated to Dreamina. (#14144)

### Runtime reliability

- Sanitize invalid tool_call arguments to unbreak strict providers.
(#14033)
- Tolerate null `function.name` in streaming tool_call deltas. (#14139)
- Preserve Gemini 3 `thoughtSignature` in `call_tools_batch`
normalization. (#14032)
- Downgrade `image_url` parts when target model lacks vision. (#14029)
- Preserve Cloudflare provider error context. (#14136)
- Use `safety_identifier` for OpenAI Responses API. (#14148)
- Unwrap underlying PG error in `formatErrorEventData`. (#14038)

---

## 🖥️ User Experience

- **Onboarding** — Preset agent naming suggestions, structured hunk ops
for `updateDocument`, persona analytics snapshot, footer promotion
pipeline, wrap-up button. (#13931, #13989, #13930, #13853, #13934)
- **Document workflow** — Agent documents promoted as primary workspace
panel; history management and compare workflow; web-crawl docs
associated with agent documents. (#13924, #13725, #13893)
- **cmdk** — Agent identity surfaced on topic search results;
topic/message search scoped to current agent. (#14204, #13960)
- **Floating chat panel** and workspace improvements. (#13887)
- **Topic completion status** with dropdown action and filter. (#14005)

---

## 🔧 Tooling

- Redis-backed feature flag provider for runtime config. (#14098)
- Vite upgraded to 8.0.0 with Rolldown strict execution order. (#12720,
#14058)
- `@lobechat/model-bank` automated npm release with provenance. (#14015,
#14017, #14018)
- Skill activation fallback when `activateTools` cannot find identifier.
(#14010)
- Cron tool: timezone and existing jobs injected into system prompt;
clarified `lobe-gtd` and `lobe-cron` descriptions. (#14012, #14013)

---

## 🔒 Security & Reliability

- **Security:** uuid bumped to v14 (advisory). (#14083)
- **Security:** validate avatar URL and scope old-avatar deletion to
owner. (#13982)
- **Security:** clear OIDC sessions on better-auth signout; return 401
(not 500) for expired OIDC JWT. (#13916, #14014)
- **Reliability:** scope pending-approval check to current assistant
turn. (#14182)
- **Reliability:** sanitize heterogeneous-agent attachment cache
filenames. (#13937)
- **Reliability:** reduce subagent task status error noise. (#14026)

---

## 👥 Contributors

Huge thanks to **17 contributors** who shipped **194 merged PRs** this
week.

@hardy · @shaun0927 · @hezhijie0327 · @sxjeru · @arvinxx · @Innei ·
@tjx666 · @lijian · @neko · @rdmclin2 · @AmAzing129 · @sudongyuer ·
@CanisMinor · @rivertwilight

Plus @lobehubbot and renovate[bot] for maintenance.

---

**Full Changelog**:
v2.1.52...v2.1.53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant