zed-industries/zed · error
Inline assistant response was interrupted ({reason:?})
Error message
Inline assistant response was interrupted ({reason:?}) What it means
Emitted when the inline assistant's completion stream stops with StopReason::MaxTokens or StopReason::Refusal before the edit completes. The response was terminated by the provider for length or policy reasons rather than finishing normally; the offending condition is the interruption stop reason on the model stream, not the tool input itself.
Source
Thrown at crates/agent_ui/src/buffer_codegen.rs:1323
return;
}
}
}
}
}
Ok(LanguageModelCompletionEvent::UsageUpdate(token_usage)) => {
*last_token_usage.lock() = token_usage;
}
Ok(LanguageModelCompletionEvent::Stop(reason)) => {
if !failure_complete {
finish_with_status(
CodegenStatus::Error(anyhow!(
"Inline assistant stopped before completing \
`{REWRITE_SECTION_TOOL_NAME}` or \
`{FAILURE_MESSAGE_TOOL_NAME}` ({reason:?})"
)),
cx,
);
} else if matches!(reason, StopReason::MaxTokens | StopReason::Refusal) {
finish_with_status(
CodegenStatus::Error(anyhow!(
"Inline assistant response was interrupted ({reason:?})"
)),
cx,
);
} else {
finish_with_status(CodegenStatus::Done, cx);
}
return;
}
Ok(LanguageModelCompletionEvent::ToolUseJsonParseError {
tool_name,
json_parse_error,
..
}) => {
finish_with_status(View on GitHub (pinned to 5a9b9558db)
Solutions
- Retry the assist, or reduce the selection size to fit the output limit
- For refusal stops, rephrase the request to avoid policy-triggering content
- Switch to a model with a larger context/output budget
Defensive patterns
Strategy: retry
When it happens
Trigger: Thrown at crates/agent_ui/src/buffer_codegen.rs:1321 when the library encounters an invalid state.
Common situations: See trigger scenarios.
AI-assisted analysis of zed-industries/zed@5a9b9558db (2026-08-20).
Data as JSON: /api/errors/f9b525cfaccd65dc.
Report an issue: GitHub.