From 98d62b9a82bfb93b960cb39966c383e2ff96d303 Mon Sep 17 00:00:00 2001 From: Mic Date: Wed, 21 May 2025 22:59:55 +0200 Subject: [PATCH] APIs list improved in the README. Fixes #382 --- README.md | 57 +++++++++++++++++++++++++++---------------------------- 1 file changed, 28 insertions(+), 29 deletions(-) diff --git a/README.md b/README.md index 989e331d..da37c349 100644 --- a/README.md +++ b/README.md @@ -17,39 +17,38 @@ Using an API integration, you can activate some automatic features:
-> [!TIP] -> **Using ChatGPT** -> -> There is no need for an API key and is possibile to use this extension even with a free account, using the **ChatGPT web interface**! -> -> If you want to connect with the OpenAI API integration, instead, you can use an **API Key**! - -
- -> [!TIP] -> **Using Google Gemini** -> -> Connect directly with Google Gemini API, using also the _System Instructions_ option if needed. - -
- -> [!TIP] -> **Using Ollama** +> [!NOTE] +> **Available Integrations** +> - **ChatGPT Web** +> - There is no need for an API key! > -> It's also possible to use a local Ollama server! +>
> -> Just remember to add `OLLAMA_ORIGINS = moz-extension://*` to the Ollama server environment variables. -> -> [More info about CORS](https://micz.it/thunderbird-addon-thunderai/ollama-cors-information/) - -
- -> [!TIP] -> **Using an OpenAI Compatible API** +> - **OpenAI API** +> - Connect directly to ChatGPT using your API key. > -> You can also use a local OpenAI Compatible API server, like LM Studio or Mistral AI! +>
> -> There is also an option to remove the "v1" segment from the API url, if needed, and to manually set the model name if the server doesn't have a models list endpoint. +> - **Google Gemini** +> - You can use also the _System Instructions_ option if needed. +> +>
+> +> - **Using Ollama** +> - Just remember to add `OLLAMA_ORIGINS = moz-extension://*` to the Ollama server environment variables. +> - [More info about CORS](https://micz.it/thunderbird-addon-thunderai/ollama-cors-information/) +> +>
+> +> - **OpenAI Compatible API** +> - You can also use a local OpenAI Compatible API server, like LM Studio or Mistral AI! +> - There is also an option to remove the "v1" segment from the API url, if needed, and to manually set the model name if the server doesn't have a models list endpoint. +> +>
+> +> - **Anthropic API** +> - Use Claude directly! +