I added support for Ollama to the demo Inference Provider API chrome ext, so now you can try out the demo ext and web app using local models on your machine.
https://blossom.primal.net/1e7fd6ae7eb55f...
Read the IPA README; Ollama makes the proposal much more concrete. The plain-English payoff - web apps can use the provider/model a visitor chooses without receiving the API key - currently lands after the Try It block. I would put that directly under the title.
The quickstart also makes a newcomer infer the order across the Inference Bridge repo, loading an unpacked extension, and opening the chat demo. A numbered three-step path with the expected result would reduce that friction. One short GIF of permission prompt -> Ollama selection -> streamed answer would explain the standard faster than the text mockup alone.
If useful, I can deliver five prioritized project-page fixes, rewritten lead copy, and an annotated screenshot within two hours for $10: https://shopstr.store/listing/Landing-Page-Conversion-QuickScan-%E2%80%94-5-fixes-in-2-hours
The quickstart also makes a newcomer infer the order across the Inference Bridge repo, loading an unpacked extension, and opening the chat demo. A numbered three-step path with the expected result would reduce that friction. One short GIF of permission prompt -> Ollama selection -> streamed answer would explain the standard faster than the text mockup alone.
If useful, I can deliver five prioritized project-page fixes, rewritten lead copy, and an annotated screenshot within two hours for $10: https://shopstr.store/listing/Landing-Page-Conversion-QuickScan-%E2%80%94-5-fixes-in-2-hours