Genkit CLI Developer UI enhanced with improved logging and error visibility
Genkit CLI `[email protected]` rolls out significant enhancements to its Developer UI, improving observability and debugging for AI applications. Developers can now view OpenTelemetry log messages directly within the UI, with inline attribute previews, and see errors labeled with their canonical status for better diagnostics. The update also brings improvements to the agent runner, including session ID support and clearer error indicators for custom patches, alongside fixes for mixed content rendering and numeric model configuration. These changes streamline the development and troubleshooting process for Genkit users, particularly those working with AI agents.
- →__Agent runner now supports a session ID in AgentInit.__
- →__Tool responses show in the prompt runner.__ Tool responses now render as tool action cards for both pending and resumed actions
Features (4) ›
__OpenTelemetry log messages are now visible in the Dev UI.__ Structured log records emitted by your app show up in the Logs panel, with each record's attributes previewed inline. Stay tunes for next JS/Go releases to start seeing actual logs
__Errors are labeled with their reported status.__ The error callout now prefixes a failed action with the canonical status the runtime returned, so a rejected request reads differently from a backend outage (for example Invalid Argument: at least one message is required in generate request)
- __Agent runner now supports a session ID in AgentInit.__
__Agent runner shows a custom patch error indicator.__ When a customPatch fails to apply, the runner surfaces an icon and tooltip, and resets state for the next turn
Fixes (4) ›
__Prose renders alongside data parts.__ A message that mixes text and data parts now renders both in order, instead of dropping the surrounding text
__Numeric model config values are sent as numbers.__ Numeric options such as temperature and topK now reach the runtime as numbers rather than strings, so strongly typed runtimes no longer reject them; stepping on float fields is fixed too
__"Open in prompt runner" works for render spans.__ It now resolves rendered prompts in agent traces, and the button is hidden when no target action can be found
- __Tool responses show in the prompt runner.__ Tool responses now render as tool action cards for both pending and resumed actions
https://github.com/genkit-ai/genkit/releases/tag/genkit-cli%401.42.0
Related releases
- Gemini for macOS: How to Use Intelligent Dictation Gemini Blog ·
- Gemini model gemini-robotics-er-1.6-preview reaches end of life in 7 days endoflife.date ·
- Explaining Google's 'Full-Stack' AI Approach Gemini Blog ·
- Genkit Python SDK v0.10.0 Adds Amazon Bedrock Support and Enterprise Persistence Genkit Releases ·
- Python GenAI Client Library Adds Audio Transcription Modes and Configuration Options Gemini Python SDK Releases ·
- Google offers free Gemini Pro/Plus for students, adds new study tools Gemini Blog ·