AIToday
Large Language ModelsZenn AI/MLPublished: Oct 17, 2026, 10:00 JST

Redmine AI Helper 3.11.0 adds web search via OpenAI, Anthropic, Gemini

Redmine AI Helper 3.11.0 adds web search via OpenAI, Anthropic, Gemini

3 Key Points

  1. What happened

    Redmine AI Helper 3.11.0 lets chat answers draw on LLM providers' web search — OpenAI, Anthropic and Gemini — with source URLs shown.

  2. Why it matters

    Teams can now get answers to questions that Redmine data alone cannot resolve, such as the cause of an error message, with sources displayed for verification, according to the release notes.

  3. What to watch

    The update requires Ruby 3.2 or higher, and SSE-based MCP servers will no longer load.

WHO IT HITSRedmine administrators and teams using the plugin for project management will gain web-search answers and easier setup checks, while those on Ruby 3.1 must upgrade before updating.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

The 3.11.0 release caps a run of updates from 3.6.1 onward that broaden what the Redmine AI Helper plugin can do. Earlier versions introduced an option to target all projects (3.7.0), project-specific ticket summary instructions (3.8.0), editable health reports (3.9.0) and a vector search connection test (3.10.0). Web search in 3.11.0 extends chat beyond Redmine's own data, while the prompt caching work addresses a structural issue: system prompts previously contained values that changed on every call, so caching barely worked. The update also tightens the technical baseline, requiring Ruby 3.2 and moving external MCP server connections to RubyLLM's built-in client.

FAQ
How do I enable web search in Redmine AI Helper?
Turn on "Enable Web Search" in the admin screen under AI Helper → General tab. It works with OpenAI (Responses API models), Anthropic and Gemini model profiles.
What are the requirements for updating to 3.11.0?
Ruby 3.2 or higher is required; Ruby 3.1 is no longer supported. External MCP server connections move to RubyLLM's built-in MCP client, and SSE-based MCP servers are not loaded.
Can I check if prompt caching is working?
Yes, if you use Langfuse, cache read and write token counts are recorded as a breakdown of input tokens.

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleClaude Opus 5.5 multi-agent tips: let managers delegate