knoten / builtin / ai / aidependencies / llm
Chrome AI
Leverage Chrome AI to use LLM directly on your device
Dieser Inhalt ist noch nicht in deiner Sprache verfügbar.
Send a prompt to the AI model built into Google Chrome (Gemini Nano) and get its reply, all on your device.
What it does
Section titled “What it does”Chrome AI is a standalone node, not a model you plug into an agent. It sends Prompt to Chrome’s built-in model and returns the answer as { result }. Nothing leaves your computer, there is no API key and no cost.
Chrome only. The node uses the AI built into Google Chrome, which Google is still rolling out (see Chrome’s built-in AI). It does not work in Firefox; there, use an on-device model from Local models or a cloud AI provider.
When to use it
Section titled “When to use it”- Short, private text jobs: rewrite a sentence, pull a name out of a paragraph, draft a one-line reply.
- Workflows that must work offline or must not send text to a cloud service.
For longer or harder tasks, a cloud model behind Basic LLM Chain gives better answers. For other on-device models, see the Local AI nodes.
Inputs and settings
Section titled “Inputs and settings”| Einstellung | Typ | Was es bewirkt |
|---|---|---|
| Prompt | Text | The prompt to use with the model. |
| Options | Group | The options to use with the model. |
| Options › Temperature | Number | The temperature to use for the model. |
| Options › Top K | Number | Sample each next token from only the K most likely candidates. Lower values are more focused; higher values are more varied (default 40). Standard: 40 |
Without options, Temperature is 0.5 and Top K is 40. Lower values give more focused, repeatable answers; higher values give more varied ones. Use expressions in Prompt to include data from earlier nodes, for example Summarise in one sentence: {{ $input.text }}.
Outputs
Section titled “Outputs”-
Output
Each output item has these fields:
| Field | Typ | What it holds |
|---|---|---|
| result | Any | The model reply; the parsed result when an output parser is connected |
{ "result": null}result holds the model’s reply as text.
Dependencies and credentials
Section titled “Dependencies and credentials”None. This node needs no connection and has no slots for other nodes.
Example workflow
Section titled “Example workflow”- A Context Menu trigger on selected text.
- Chrome AI with Prompt
Rewrite this politely: {{ $input.selectionText }}. - Clipboard Write copies
{{ $input.result }}.
Common issues
Section titled “Common issues”- It fails with “only works in Chrome”. You are in Firefox, which has no built-in AI. Switch to an on-device or cloud model.
- The model is not available. Chrome’s built-in AI is not enabled or not downloaded yet on this computer. Follow Google’s setup steps for Chrome’s built-in AI, then try again.
- The run was stopped and there is no reply. Stop cancels the answer while the model is still writing.
Related nodes
Section titled “Related nodes”- Basic LLM Chain: prompt a cloud or local chat model.
- Generate Text: on-device text generation in any browser.
- Summarize: on-device summaries.