fix(openai): send content: null instead of "" for tool-call-only assistant messages #12390
+78
−2
Conversation
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
…only assistant messages Per the OpenAI API spec, `content` on assistant messages is nullable. When an assistant turn only contains tool calls and no text, the content field should be `null` rather than an empty string. Sending `content: ""` causes issues with OpenAI-compatible providers (e.g. Ollama) where the empty string is rendered differently by model templates, causing models like qwen3-coder to switch from structured tool_calls to text-based markup on subsequent turns. Fixes vercel#12389 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
4 tasks
lbijeau
added a commit
to lbijeau/meloqui
that referenced
this pull request
Feb 10, 2026
… SDK and fix content hygiene Replaces the native Ollama tool-calling loop with Vercel AI SDK's generateText/streamText, adds a fetch-level workaround for the empty-content bug that causes qwen3-coder to drop structured tool_calls, and introduces streaming code-block re-prompt detection for models that output tools as markdown. Upstream fixes pending: - vercel/ai#12390 (SDK sends content:"" instead of null) - ollama/ollama#14182 (Ollama /v1 layer doesn't normalise "") Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.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
convertToOpenAIChatMessagesnow sendscontent: nullinstead ofcontent: ""for assistant messages that only contain tool calls and no text contentcontenton assistant messages is nullable — their examples showcontent: nullwhen the assistant only makes tool callsChatCompletionAssistantMessage) already declaredcontent?: string | null, so this aligns the runtime behavior with the typeMotivation
Sending
content: ""causes issues with OpenAI-compatible providers (e.g. Ollama) where the empty string is rendered differently by model templates. Specifically, with Ollama + qwen3-coder, the empty string in prior assistant messages causes the model to switch from structuredtool_callsto text-based markup (<function=name>) on subsequent turns, breaking multi-turn tool calling.Verified via curl testing that the same conversation with
content: nullkeeps the model in structured tool-calling mode, whilecontent: ""causes mode switching.Also filed on the Ollama side: ollama/ollama#14181
Test plan
content: nullcontent: nullcontentconvert-to-openai-chat-messages.test.tspassFixes #12389