Merge pull request #1875 from hlohaus/carst

Add image model list
2024-12-24 11:34:02 +03:00 · 2024-04-22 01:35:07 +02:00 · 2024-04-22 01:35:07 +02:00 · 4b4d1f08b5
commit 4b4d1f08b5
parent 0f3935f1c0 4bc4d635bc
33 changed files with 553 additions and 312 deletions
--- a/.gitignore
+++ b/.gitignore
@ -55,6 +55,7 @@ local.py
 image.py
 .buildozer
 hardir
+har_and_cookies
 node_modules
 models
 projects/windows/g4f
--- a/README.md
+++ b/README.md
@ -91,7 +91,7 @@ As per the survey, here is a list of improvements to come

 ```sh
 docker pull hlohaus789/g4f
-docker run -p 8080:8080 -p 1337:1337 -p 7900:7900 --shm-size="2g" -v ${PWD}/hardir:/app/hardir hlohaus789/g4f:latest
+docker run -p 8080:8080 -p 1337:1337 -p 7900:7900 --shm-size="2g" -v ${PWD}/har_and_cookies:/app/har_and_cookies hlohaus789/g4f:latest
 ```

 3. **Access the Client:** 
@ -217,10 +217,11 @@ Access with: http://localhost:1337/v1

 #### Cookies

-You need cookies for BingCreateImages and the Gemini Provider.
-From Bing you need the "_U" cookie and from Gemini you need the "__Secure-1PSID" cookie.
-Sometimes you doesn't need the "__Secure-1PSID" cookie, but some other auth cookies.
-You can pass the cookies in the create function or you use the `set_cookies` setter before you run G4F:
+Cookies are essential for using Meta AI and Microsoft Designer to create images.
+Additionally, cookies are required for the Google Gemini and WhiteRabbitNeo Provider.
+From Bing, ensure you have the "_U" cookie, and from Google, all cookies starting with "__Secure-1PSID" are needed.
+
+You can pass these cookies directly to the create function or set them using the `set_cookies` method before running G4F:

 ```python
 from g4f.cookies import set_cookies
@ -228,10 +229,25 @@ from g4f.cookies import set_cookies
 set_cookies(".bing.com", {
  "_U": "cookie value"
 })
+
 set_cookies(".google.com", {
  "__Secure-1PSID": "cookie value"
 })
-...
+```
+
+Alternatively, you can place your .har and cookie files in the `/har_and_cookies` directory. To export a cookie file, use the EditThisCookie extension available on the Chrome Web Store: [EditThisCookie Extension](https://chromewebstore.google.com/detail/editthiscookie/fngmhnnpilhplaeedifhccceomclgfbg).
+
+You can also create .har files to capture cookies. If you need further assistance, refer to the next section.
+
+```bash
+python -m g4f.cli api --debug
+```
+```
+Read .har file: ./har_and_cookies/you.com.har
+Cookies added: 10 from .you.com
+Read cookie file: ./har_and_cookies/google.json
+Cookies added: 16 from .google.com
+Starting server... [g4f v-0.0.0] (debug)
 ```

 #### .HAR File for OpenaiChat Provider
@ -249,7 +265,7 @@ To utilize the OpenaiChat provider, a .har file is required from https://chat.op

 ##### Storing the .HAR File

- Place the exported .har file in the `./hardir` directory if you are using Docker. Alternatively, you can store it in any preferred location within your current working directory.
+- Place the exported .har file in the `./har_and_cookies` directory if you are using Docker. Alternatively, you can store it in any preferred location within your current working directory.

 Note: Ensure that your .har file is stored securely, as it may contain sensitive information.

@ -273,13 +289,13 @@ set G4F_PROXY=http://host:port

 | Website | Provider | GPT-3.5 | GPT-4 | Stream | Status | Auth |
 | ------  | -------  | ------- | ----- | ------ | ------ | ---- |
-| [bing.com](https://bing.com/chat) | `g4f.Provider.Bing` | ❌ | ✔️ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
-| [chatgpt.ai](https://chatgpt.ai) | `g4f.Provider.ChatgptAi` | ❌ | ✔️ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [liaobots.site](https://liaobots.site) | `g4f.Provider.Liaobots` | ✔️ | ✔️ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [chat.openai.com](https://chat.openai.com) | `g4f.Provider.OpenaiChat` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ✔️ |
+| [bing.com](https://bing.com/chat) | `g4f.Provider.Bing` | ❌ | ✔️ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [chatgpt.ai](https://chatgpt.ai) | `g4f.Provider.ChatgptAi` | ❌ | ✔️ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [liaobots.site](https://liaobots.site) | `g4f.Provider.Liaobots` | ✔️ | ✔️ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [chat.openai.com](https://chat.openai.com) | `g4f.Provider.OpenaiChat` | ✔️ | ✔️ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌+✔️ |
 | [raycast.com](https://raycast.com) | `g4f.Provider.Raycast` | ✔️ | ✔️ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ✔️ |
 | [beta.theb.ai](https://beta.theb.ai) | `g4f.Provider.Theb` | ✔️ | ✔️ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
-| [you.com](https://you.com) | `g4f.Provider.You` | ✔️ | ✔️ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [you.com](https://you.com) | `g4f.Provider.You` | ✔️ | ✔️ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |

 ## Best OpenSource Models
 While we wait for gpt-5, here is a list of new models that are at least better than gpt-3.5-turbo. **Some are better than gpt-4**. Expect this list to grow.
@ -296,19 +312,24 @@ While we wait for gpt-5, here is a list of new models that are at least better t
 | Website | Provider | GPT-3.5 | GPT-4 | Stream | Status | Auth |
 | ------  | -------  | ------- | ----- | ------ | ------ | ---- |
 | [chat3.aiyunos.top](https://chat3.aiyunos.top/) | `g4f.Provider.AItianhuSpace` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
-| [chatforai.store](https://chatforai.store) | `g4f.Provider.ChatForAi` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [chatgpt4online.org](https://chatgpt4online.org) | `g4f.Provider.Chatgpt4Online` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [chatgpt-free.cc](https://www.chatgpt-free.cc) | `g4f.Provider.ChatgptNext` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [chatgptx.de](https://chatgptx.de) | `g4f.Provider.ChatgptX` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [flowgpt.com](https://flowgpt.com/chat) | `g4f.Provider.FlowGpt` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [freegptsnav.aifree.site](https://freegptsnav.aifree.site) | `g4f.Provider.FreeGpt` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
-| [gpttalk.ru](https://gpttalk.ru) | `g4f.Provider.GptTalkRu` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [koala.sh](https://koala.sh) | `g4f.Provider.Koala` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [chat10.aichatos.xyz](https://chat10.aichatos.xyz) | `g4f.Provider.Aichatos` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [chatforai.store](https://chatforai.store) | `g4f.Provider.ChatForAi` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [chatgpt4online.org](https://chatgpt4online.org) | `g4f.Provider.Chatgpt4Online` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [chatgpt-free.cc](https://www.chatgpt-free.cc) | `g4f.Provider.ChatgptNext` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [chatgptx.de](https://chatgptx.de) | `g4f.Provider.ChatgptX` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [f1.cnote.top](https://f1.cnote.top) | `g4f.Provider.Cnote` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [duckduckgo.com](https://duckduckgo.com/duckchat) | `g4f.Provider.DuckDuckGo` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [ecosia.org](https://www.ecosia.org) | `g4f.Provider.Ecosia` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [feedough.com](https://www.feedough.com) | `g4f.Provider.Feedough` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [flowgpt.com](https://flowgpt.com/chat) | `g4f.Provider.FlowGpt` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [freegptsnav.aifree.site](https://freegptsnav.aifree.site) | `g4f.Provider.FreeGpt` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [gpttalk.ru](https://gpttalk.ru) | `g4f.Provider.GptTalkRu` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [koala.sh](https://koala.sh) | `g4f.Provider.Koala` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
 | [app.myshell.ai](https://app.myshell.ai/chat) | `g4f.Provider.MyShell` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
 | [perplexity.ai](https://www.perplexity.ai) | `g4f.Provider.PerplexityAi` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
 | [poe.com](https://poe.com) | `g4f.Provider.Poe` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ✔️ |
 | [talkai.info](https://talkai.info) | `g4f.Provider.TalkAi` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
-| [chat.vercel.ai](https://chat.vercel.ai) | `g4f.Provider.Vercel` | ✔️ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [chat.vercel.ai](https://chat.vercel.ai) | `g4f.Provider.Vercel` | ✔️ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
 | [aitianhu.com](https://www.aitianhu.com) | `g4f.Provider.AItianhu` | ✔️ | ❌ | ✔️ | ![Inactive](https://img.shields.io/badge/Inactive-red) | ❌ |
 | [chatgpt.bestim.org](https://chatgpt.bestim.org) | `g4f.Provider.Bestim` | ✔️ | ❌ | ✔️ | ![Inactive](https://img.shields.io/badge/Inactive-red) | ❌ |
 | [chatbase.co](https://www.chatbase.co) | `g4f.Provider.ChatBase` | ✔️ | ❌ | ✔️ | ![Inactive](https://img.shields.io/badge/Inactive-red) | ❌ |
@ -328,48 +349,68 @@ While we wait for gpt-5, here is a list of new models that are at least better t

 | Website | Provider | GPT-3.5 | GPT-4 | Stream | Status | Auth |
 | ------  | -------  | ------- | ----- | ------ | ------ | ---- |
-| [openchat.team](https://openchat.team) | `g4f.Provider.Aura` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [bard.google.com](https://bard.google.com) | `g4f.Provider.Bard` | ❌ | ❌ | ❌ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ✔️ |
+| [openchat.team](https://openchat.team) | `g4f.Provider.Aura` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [blackbox.ai](https://www.blackbox.ai) | `g4f.Provider.Blackbox` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [cohereforai-c4ai-command-r-plus.hf.space](https://cohereforai-c4ai-command-r-plus.hf.space) | `g4f.Provider.Cohere` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
 | [deepinfra.com](https://deepinfra.com) | `g4f.Provider.DeepInfra` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [free.chatgpt.org.uk](https://free.chatgpt.org.uk) | `g4f.Provider.FreeChatgpt` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [free.chatgpt.org.uk](https://free.chatgpt.org.uk) | `g4f.Provider.FreeChatgpt` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
 | [gemini.google.com](https://gemini.google.com) | `g4f.Provider.Gemini` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ✔️ |
 | [ai.google.dev](https://ai.google.dev) | `g4f.Provider.GeminiPro` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ✔️ |
 | [gemini-chatbot-sigma.vercel.app](https://gemini-chatbot-sigma.vercel.app) | `g4f.Provider.GeminiProChat` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [developers.sber.ru](https://developers.sber.ru/gigachat) | `g4f.Provider.GigaChat` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ✔️ |
+| [console.groq.com](https://console.groq.com/playground) | `g4f.Provider.Groq` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ✔️ |
 | [huggingface.co](https://huggingface.co/chat) | `g4f.Provider.HuggingChat` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
 | [huggingface.co](https://huggingface.co/chat) | `g4f.Provider.HuggingFace` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [llama2.ai](https://www.llama2.ai) | `g4f.Provider.Llama2` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
+| [llama2.ai](https://www.llama2.ai) | `g4f.Provider.Llama` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [meta.ai](https://www.meta.ai) | `g4f.Provider.MetaAI` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌+✔️ |
+| [openrouter.ai](https://openrouter.ai) | `g4f.Provider.OpenRouter` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ✔️ |
 | [labs.perplexity.ai](https://labs.perplexity.ai) | `g4f.Provider.PerplexityLabs` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [pi.ai](https://pi.ai/talk) | `g4f.Provider.Pi` | ❌ | ❌ | ✔️ | ![Active](https://img.shields.io/badge/Active-brightgreen) | ❌ |
-| [theb.ai](https://theb.ai) | `g4f.Provider.ThebApi` | ❌ | ❌ | ❌ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ✔️ |
-| [open-assistant.io](https://open-assistant.io/chat) | `g4f.Provider.OpenAssistant` | ❌ | ❌ | ✔️ | ![Inactive](https://img.shields.io/badge/Inactive-red) | ✔️ |
+| [pi.ai](https://pi.ai/talk) | `g4f.Provider.Pi` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [replicate.com](https://replicate.com) | `g4f.Provider.ReplicateImage` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ❌ |
+| [theb.ai](https://theb.ai) | `g4f.Provider.ThebApi` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ✔️ |
+| [whiterabbitneo.com](https://www.whiterabbitneo.com) | `g4f.Provider.WhiteRabbitNeo` | ❌ | ❌ | ✔️ | ![Unknown](https://img.shields.io/badge/Unknown-grey) | ✔️ |
+| [bard.google.com](https://bard.google.com) | `g4f.Provider.Bard` | ❌ | ❌ | ❌ | ![Inactive](https://img.shields.io/badge/Inactive-red) | ✔️ |

 ### Models

-| Model                       | Base Provider | Provider | Website |
-|-----------------------------| ------------- | -------- | ------- |
-| gpt-3.5-turbo               | OpenAI | 5+ Providers | [openai.com](https://openai.com/) |
-| gpt-4                       | OpenAI | 2+ Providers | [openai.com](https://openai.com/) |
-| gpt-4-turbo                 | OpenAI | g4f.Provider.Bing | [openai.com](https://openai.com/) |
-| Llama-2-7b-chat-hf          | Meta | 2+ Providers | [llama.meta.com](https://llama.meta.com/) |
-| Llama-2-13b-chat-hf         | Meta | 2+ Providers | [llama.meta.com](https://llama.meta.com/) |
-| Llama-2-70b-chat-hf         | Meta | 3+ Providers | [llama.meta.com](https://llama.meta.com/) |
-| Meta-Llama-3-8b             | Meta | 3+ Providers | [llama.meta.com](https://llama.meta.com/) |
-| Meta-Llama-3-70b            | Meta | 3+ Providers | [llama.meta.com](https://llama.meta.com/) |
-| CodeLlama-34b-Instruct-hf   | Meta | 2+ Providers | [llama.meta.com](https://llama.meta.com/) |
-| CodeLlama-70b-Instruct-hf   | Meta | 2+ Providers | [llama.meta.com](https://llama.meta.com/) |
-| Mixtral-8x7B-Instruct-v0.1  | Huggingface | 4+ Providers | [huggingface.co](https://huggingface.co/) |
-| Mistral-7B-Instruct-v0.1    | Huggingface | 4+ Providers | [huggingface.co](https://huggingface.co/) |
-| dolphin-2.6-mixtral-8x7b    | Huggingface | g4f.Provider.DeepInfra | [huggingface.co](https://huggingface.co/) |
-| lzlv_70b_fp16_hf            | Huggingface | g4f.Provider.DeepInfra | [huggingface.co](https://huggingface.co/) |
-| airoboros-70b               | Huggingface | g4f.Provider.DeepInfra | [huggingface.co](https://huggingface.co/) |
-| airoboros-l2-70b-gpt4-1.4.1 | Huggingface | g4f.Provider.DeepInfra | [huggingface.co](https://huggingface.co/) |
-| openchat_3.5                | Huggingface | 2+ Providers | [huggingface.co](https://huggingface.co/) |
-| gemini                      | Google | g4f.Provider.Gemini | [gemini.google.com](https://gemini.google.com/) |
-| gemini-pro                  | Google | 2+ Providers | [gemini.google.com](https://gemini.google.com/) |
-| claude-v2                   | Anthropic | 1+ Providers | [anthropic.com](https://www.anthropic.com/) |
-| claude-3-opus               | Anthropic | g4f.Provider.You | [anthropic.com](https://www.anthropic.com/) |
-| claude-3-sonnet             | Anthropic | g4f.Provider.You | [anthropic.com](https://www.anthropic.com/) |
-| pi                          | Inflection | g4f.Provider.Pi | [inflection.ai](https://inflection.ai/) |
+| Model | Base Provider | Provider | Website |
+| ----- | ------------- | -------- | ------- |
+| gpt-3.5-turbo | OpenAI | 8+ Providers | [openai.com](https://openai.com/) |
+| gpt-4 | OpenAI | 2+ Providers | [openai.com](https://openai.com/) |
+| gpt-4-turbo | OpenAI | g4f.Provider.Bing | [openai.com](https://openai.com/) |
+| Llama-2-7b-chat-hf | Meta | 2+ Providers | [llama.meta.com](https://llama.meta.com/) |
+| Llama-2-13b-chat-hf | Meta | 2+ Providers | [llama.meta.com](https://llama.meta.com/) |
+| Llama-2-70b-chat-hf | Meta | 3+ Providers | [llama.meta.com](https://llama.meta.com/) |
+| Meta-Llama-3-8b-instruct | Meta | 1+ Providers | [llama.meta.com](https://llama.meta.com/) |
+| Meta-Llama-3-70b-instruct | Meta | 2+ Providers | [llama.meta.com](https://llama.meta.com/) |
+| CodeLlama-34b-Instruct-hf | Meta | g4f.Provider.HuggingChat | [llama.meta.com](https://llama.meta.com/) |
+| CodeLlama-70b-Instruct-hf | Meta | 2+ Providers | [llama.meta.com](https://llama.meta.com/) |
+| Mixtral-8x7B-Instruct-v0.1 | Huggingface | 4+ Providers | [huggingface.co](https://huggingface.co/) |
+| Mistral-7B-Instruct-v0.1 | Huggingface | 3+ Providers | [huggingface.co](https://huggingface.co/) |
+| Mistral-7B-Instruct-v0.2 | Huggingface | g4f.Provider.DeepInfra | [huggingface.co](https://huggingface.co/) |
+| zephyr-orpo-141b-A35b-v0.1 | Huggingface | 2+ Providers | [huggingface.co](https://huggingface.co/) |
+| dolphin-2.6-mixtral-8x7b | Huggingface | g4f.Provider.DeepInfra | [huggingface.co](https://huggingface.co/) |
+| gemini | Google | g4f.Provider.Gemini | [gemini.google.com](https://gemini.google.com/) |
+| gemini-pro | Google | 2+ Providers | [gemini.google.com](https://gemini.google.com/) |
+| claude-v2 | Anthropic | 1+ Providers | [anthropic.com](https://www.anthropic.com/) |
+| claude-3-opus | Anthropic | g4f.Provider.You | [anthropic.com](https://www.anthropic.com/) |
+| claude-3-sonnet | Anthropic | g4f.Provider.You | [anthropic.com](https://www.anthropic.com/) |
+| lzlv_70b_fp16_hf | Huggingface | g4f.Provider.DeepInfra | [huggingface.co](https://huggingface.co/) |
+| airoboros-70b | Huggingface | g4f.Provider.DeepInfra | [huggingface.co](https://huggingface.co/) |
+| openchat_3.5 | Huggingface | 2+ Providers | [huggingface.co](https://huggingface.co/) |
+| pi | Inflection | g4f.Provider.Pi | [inflection.ai](https://inflection.ai/) |
+
+### Image and Vision Models
+
+| Label | Provider | Image Model | Vision Model | Website |
+| ----- | -------- | ----------- | ------------ | ------- |
+| Microsoft Copilot in Bing | `g4f.Provider.Bing` | dall-e| gpt-4-vision | [bing.com](https://bing.com/chat) |
+| DeepInfra | `g4f.Provider.DeepInfra` | stability-ai/sdxl| llava-1.5-7b-hf | [deepinfra.com](https://deepinfra.com) |
+| Gemini | `g4f.Provider.Gemini` | gemini| gemini | [gemini.google.com](https://gemini.google.com) |
+| Meta AI | `g4f.Provider.MetaAI` | meta| ❌ | [meta.ai](https://www.meta.ai) |
+| OpenAI ChatGPT | `g4f.Provider.OpenaiChat` | dall-e| gpt-4-vision | [chat.openai.com](https://chat.openai.com) |
+| Replicate | `g4f.Provider.Replicate` | stability-ai/sdxl| ❌ | [replicate.com](https://replicate.com) |
+| You.com | `g4f.Provider.You` | dall-e| agent | [you.com](https://you.com) |

 ## 🔗 Powered by gpt4free

--- a/etc/tool/readme_table.py
+++ b/etc/tool/readme_table.py
@ -14,6 +14,8 @@ async def test_async(provider: ProviderType):
        return False
    messages = [{"role": "user", "content": "Hello Assistant!"}]
    try:
+        if "webdriver" in provider.get_parameters():
+            return False
        response = await asyncio.wait_for(ChatCompletion.create_async(
            model=models.default,
            messages=messages,
@ -88,7 +90,7 @@ def print_models():
        "huggingface": "Huggingface",
        "anthropic": "Anthropic",
        "inflection": "Inflection",
-        "meta": "Meta"
+        "meta": "Meta",
    }
    provider_urls = {
        "google": "https://gemini.google.com/",
@ -96,7 +98,7 @@ def print_models():
        "huggingface": "https://huggingface.co/",
        "anthropic": "https://www.anthropic.com/",
        "inflection": "https://inflection.ai/",
-        "meta": "https://llama.meta.com/"
+        "meta": "https://llama.meta.com/",
    }

    lines = [
@ -108,6 +110,8 @@ def print_models():
            if name not in ("gpt-3.5-turbo", "gpt-4", "gpt-4-turbo"):
                continue
        name = re.split(r":|/", model.name)[-1]
+        if model.base_provider not in base_provider_names:
+            continue
        base_provider = base_provider_names[model.base_provider]
        if not isinstance(model.best_provider, BaseRetryProvider):
            provider_name = f"g4f.Provider.{model.best_provider.__name__}"
@ -121,7 +125,26 @@ def print_models():

    print("\n".join(lines))

+def print_image_models():
+    lines = [
+        "| Label | Provider | Image Model | Vision Model | Website |",
+        "| ----- | -------- | ----------- | ------------ | ------- |",
+    ]
+    from g4f.gui.server.api import Api
+    for image_model in Api.get_image_models():
+        provider_url = image_model["url"]
+        netloc = urlparse(provider_url).netloc.replace("www.", "")
+        website = f"[{netloc}]({provider_url})"
+        label = image_model["provider"] if image_model["label"] is None else image_model["label"]
+        if image_model["vision_model"] is None:
+            image_model["vision_model"] = "❌"
+        lines.append(f'| {label} | `g4f.Provider.{image_model["provider"]}` | {image_model["image_model"]}| {image_model["vision_model"]} | {website} |')
+
+    print("\n".join(lines))
+
 if __name__ == "__main__":
-    print_providers()
+    #print_providers()
+    #print("\n", "-" * 50, "\n")
+    #print_models()
    print("\n", "-" * 50, "\n")
-    print_models()
+    print_image_models()
--- a/g4f/Provider/Bing.py
+++ b/g4f/Provider/Bing.py
@ -38,8 +38,9 @@ class Bing(AsyncGeneratorProvider, ProviderModelMixin):
    supports_message_history = True
    supports_gpt_4 = True
    default_model = "Balanced"
+    default_vision_model = "gpt-4-vision"
    models = [getattr(Tones, key) for key in Tones.__dict__ if not key.startswith("__")]
-        
+
    @classmethod
    def create_async_generator(
        cls,
--- a/g4f/Provider/BingCreateImages.py
+++ b/g4f/Provider/BingCreateImages.py
@ -13,9 +13,11 @@ from .bing.create_images import create_images, create_session, get_cookies_from_

 class BingCreateImages(AsyncGeneratorProvider, ProviderModelMixin):
    label = "Microsoft Designer"
+    parent = "Bing"
    url = "https://www.bing.com/images/create"
    working = True
    needs_auth = True
+    image_models = ["dall-e"]

    def __init__(self, cookies: Cookies = None, proxy: str = None) -> None:
        self.cookies: Cookies = cookies
--- a/g4f/Provider/DeepInfra.py
+++ b/g4f/Provider/DeepInfra.py
@ -1,17 +1,22 @@
 from __future__ import annotations

 import requests
-from ..typing import AsyncResult, Messages
+from ..typing import AsyncResult, Messages, ImageType
+from ..image import to_data_uri
 from .needs_auth.Openai import Openai

 class DeepInfra(Openai):
    label = "DeepInfra"
    url = "https://deepinfra.com"
    working = True
-    needs_auth = False
+    has_auth = True
    supports_stream = True
    supports_message_history = True
-    default_model = 'HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1'
+    default_model = "meta-llama/Meta-Llama-3-70b-instruct"
+    default_vision_model = "llava-hf/llava-1.5-7b-hf"
+    model_aliases = {
+        'mixtral-8x22b': 'HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1'
+    }

    @classmethod
    def get_models(cls):
@ -27,19 +32,12 @@ class DeepInfra(Openai):
        model: str,
        messages: Messages,
        stream: bool,
+        image: ImageType = None,
        api_base: str = "https://api.deepinfra.com/v1/openai",
        temperature: float = 0.7,
        max_tokens: int = 1028,
        **kwargs
    ) -> AsyncResult:
-        
-        if not '/' in model:
-            models = {
-                'mixtral-8x22b': 'HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1',
-                'dbrx-instruct': 'databricks/dbrx-instruct',
-            }
-            model = models.get(model, model)
-        
        headers = {
            'Accept-Encoding': 'gzip, deflate, br',
            'Accept-Language': 'en-US',
@ -55,6 +53,19 @@ class DeepInfra(Openai):
            'sec-ch-ua-mobile': '?0',
            'sec-ch-ua-platform': '"macOS"',
        }
+        if image is not None:
+            if not model:
+                model = cls.default_vision_model
+            messages[-1]["content"] = [
+                {
+                    "type": "image_url",
+                    "image_url": {"url": to_data_uri(image)}
+                },
+                {
+                    "type": "text",
+                    "text": messages[-1]["content"]
+                }
+            ]
        return super().create_async_generator(
            model, messages,
            stream=stream,
--- a/g4f/Provider/DeepInfraImage.py
+++ b/g4f/Provider/DeepInfraImage.py
@ -9,8 +9,10 @@ from ..image import ImageResponse

 class DeepInfraImage(AsyncGeneratorProvider, ProviderModelMixin):
    url = "https://deepinfra.com"
+    parent = "DeepInfra"
    working = True
    default_model = 'stability-ai/sdxl'
+    image_models = [default_model]

    @classmethod
    def get_models(cls):
@ -18,6 +20,7 @@ class DeepInfraImage(AsyncGeneratorProvider, ProviderModelMixin):
            url = 'https://api.deepinfra.com/models/featured'
            models = requests.get(url).json()
            cls.models = [model['model_name'] for model in models if model["reported_type"] == "text-to-image"]
+            cls.image_models = cls.models
        return cls.models

    @classmethod
--- a/g4f/Provider/MetaAI.py
+++ b/g4f/Provider/MetaAI.py
@ -13,7 +13,7 @@ from ..requests import raise_for_status, DEFAULT_HEADERS
 from ..image import ImageResponse, ImagePreview
 from ..errors import ResponseError
 from .base_provider import AsyncGeneratorProvider
-from .helper import format_prompt, get_connector
+from .helper import format_prompt, get_connector, format_cookies

 class Sources():
    def __init__(self, link_list: List[Dict[str, str]]) -> None:
@ -48,7 +48,6 @@ class MetaAI(AsyncGeneratorProvider):

    async def update_access_token(self, birthday: str = "1999-01-01"):
        url = "https://www.meta.ai/api/graphql/"
-
        payload = {
            "lsd": self.lsd,
            "fb_api_caller_class": "RelayModern",
@ -90,7 +89,7 @@ class MetaAI(AsyncGeneratorProvider):
            headers = {}
        headers = {
            'content-type': 'application/x-www-form-urlencoded',
-            'cookie': "; ".join([f"{k}={v}" for k, v in cookies.items()]),
+            'cookie': format_cookies(self.cookies),
            'origin': 'https://www.meta.ai',
            'referer': 'https://www.meta.ai/',
            'x-asbd-id': '129477',
@ -194,7 +193,7 @@ class MetaAI(AsyncGeneratorProvider):
            **headers
        }
        async with self.session.post(url, headers=headers, cookies=self.cookies, data=payload) as response:
-            await raise_for_status(response)
+            await raise_for_status(response, "Fetch sources failed")
            text = await response.text()
            if "<h1>Something Went Wrong</h1>" in text:
                raise ResponseError("Response: Something Went Wrong")
--- a/g4f/Provider/MetaAIAccount.py
+++ b/g4f/Provider/MetaAIAccount.py
@ -6,6 +6,8 @@ from .MetaAI import MetaAI

 class MetaAIAccount(MetaAI):
    needs_auth = True
+    parent = "MetaAI"
+    image_models = ["meta"]

    @classmethod
    async def create_async_generator(
--- a/g4f/Provider/Replicate.py
+++ b/g4f/Provider/Replicate.py
@ -0,0 +1,84 @@
+from __future__ import annotations
+
+from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
+from .helper import format_prompt, filter_none
+from ..typing import AsyncResult, Messages
+from ..requests import raise_for_status
+from ..requests.aiohttp import StreamSession
+from ..errors import ResponseError, MissingAuthError
+
+class Replicate(AsyncGeneratorProvider, ProviderModelMixin):
+    url = "https://replicate.com"
+    working = True
+    default_model = "meta/meta-llama-3-70b-instruct"
+
+    @classmethod
+    async def create_async_generator(
+        cls,
+        model: str,
+        messages: Messages,
+        api_key: str = None,
+        proxy: str = None,
+        timeout: int = 180,
+        system_prompt: str = None,
+        max_new_tokens: int = None,
+        temperature: float = None,
+        top_p: float = None,
+        top_k: float = None,
+        stop: list = None,
+        extra_data: dict = {},
+        headers: dict = {
+            "accept": "application/json",
+        },
+        **kwargs
+    ) -> AsyncResult:
+        model = cls.get_model(model)
+        if cls.needs_auth and api_key is None:
+            raise MissingAuthError("api_key is missing")
+        if api_key is not None:
+            headers["Authorization"] = f"Bearer {api_key}"
+            api_base = "https://api.replicate.com/v1/models/"
+        else:
+            api_base = "https://replicate.com/api/models/"
+        async with StreamSession(
+            proxy=proxy,
+            headers=headers,
+            timeout=timeout
+        ) as session:
+            data = {
+                "stream": True,
+                "input": {
+                    "prompt": format_prompt(messages),
+                    **filter_none(
+                        system_prompt=system_prompt,
+                        max_new_tokens=max_new_tokens,
+                        temperature=temperature,
+                        top_p=top_p,
+                        top_k=top_k,
+                        stop_sequences=",".join(stop) if stop else None
+                    ),
+                    **extra_data
+                },
+            }
+            url = f"{api_base.rstrip('/')}/{model}/predictions"
+            async with session.post(url, json=data) as response:
+                message = "Model not found" if response.status == 404 else None
+                await raise_for_status(response, message)
+                result = await response.json()
+                if "id" not in result:
+                    raise ResponseError(f"Invalid response: {result}")
+                async with session.get(result["urls"]["stream"], headers={"Accept": "text/event-stream"}) as response:
+                    await raise_for_status(response)
+                    event = None
+                    async for line in response.iter_lines():
+                        if line.startswith(b"event: "):
+                            event = line[7:]
+                            if event == b"done":
+                                break
+                        elif event == b"output":
+                            if line.startswith(b"data: "):
+                                new_text = line[6:].decode()
+                                if new_text:
+                                    yield new_text
+                                else:
+                                    yield "\n"
--- a/g4f/Provider/ReplicateImage.py
+++ b/g4f/Provider/ReplicateImage.py
@ -11,12 +11,14 @@ from ..errors import ResponseError

 class ReplicateImage(AsyncGeneratorProvider, ProviderModelMixin):
    url = "https://replicate.com"
+    parent = "Replicate"
    working = True
    default_model = 'stability-ai/sdxl'
    default_versions = [
        "39ed52f2a78e934b3ba6e2a89f5b1c712de7dfea535525255b1aa35c5565e08b",
        "2b017d9b67edd2ee1401238df49d75da53c523f36e363881e057f5dc3ed3c5b2"
    ]
+    image_models = [default_model]

    @classmethod
    async def create_async_generator(
--- a/g4f/Provider/You.py
+++ b/g4f/Provider/You.py
@ -8,19 +8,22 @@ import uuid
 from ..typing import AsyncResult, Messages, ImageType, Cookies
 from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
 from .helper import format_prompt
-from ..image import ImageResponse, to_bytes, is_accepted_format
+from ..image import ImageResponse, ImagePreview, to_bytes, is_accepted_format
 from ..requests import StreamSession, FormData, raise_for_status
 from .you.har_file import get_telemetry_ids
 from .. import debug

 class You(AsyncGeneratorProvider, ProviderModelMixin):
+    label = "You.com"
    url = "https://you.com"
    working = True
    supports_gpt_35_turbo = True
    supports_gpt_4 = True
    default_model = "gpt-3.5-turbo"
+    default_vision_model = "agent"
+    image_models = ["dall-e"]
    models = [
-        "gpt-3.5-turbo",
+        default_model,
        "gpt-4",
        "gpt-4-turbo",
        "claude-instant",
@ -29,7 +32,8 @@ class You(AsyncGeneratorProvider, ProviderModelMixin):
        "claude-3-sonnet",
        "gemini-pro",
        "zephyr",
-        "dall-e",
+        default_vision_model,
+        *image_models
    ]
    model_aliases = {
        "claude-v2": "claude-2"
@ -51,7 +55,7 @@ class You(AsyncGeneratorProvider, ProviderModelMixin):
        chat_mode: str = "default",
        **kwargs,
    ) -> AsyncResult:
-        if image is not None:
+        if image is not None or model == cls.default_vision_model:
            chat_mode = "agent"
        elif not model or model == cls.default_model:
            ...
@ -62,13 +66,18 @@ class You(AsyncGeneratorProvider, ProviderModelMixin):
            chat_mode = "custom"
            model = cls.get_model(model)
        async with StreamSession(
-            proxies={"all": proxy},
+            proxy=proxy,
            impersonate="chrome",
            timeout=(30, timeout)
        ) as session:
            cookies = await cls.get_cookies(session) if chat_mode != "default" else None
-            
-            upload = json.dumps([await cls.upload_file(session, cookies, to_bytes(image), image_name)]) if image else ""
+            upload = ""
+            if image is not None:
+                upload_file = await cls.upload_file(
+                    session, cookies,
+                    to_bytes(image), image_name
+                )
+                upload = json.dumps([upload_file])
            headers = {
                "Accept": "text/event-stream",
                "Referer": f"{cls.url}/search?fromSearchBar=true&tbm=youchat",
@ -102,11 +111,17 @@ class You(AsyncGeneratorProvider, ProviderModelMixin):
                        if event == "youChatToken" and event in data:
                            yield data[event]
                        elif event == "youChatUpdate" and "t" in data and data["t"] is not None:
-                            match = re.search(r"!\[fig\]\((.+?)\)", data["t"])
-                            if match:
-                                yield ImageResponse(match.group(1), messages[-1]["content"])
+                            if chat_mode == "create":
+                                match = re.search(r"!\[(.+?)\]\((.+?)\)", data["t"])
+                                if match:
+                                    if match.group(1) == "fig":
+                                        yield ImagePreview(match.group(2), messages[-1]["content"])
+                                    else:
+                                        yield ImageResponse(match.group(2), match.group(1))
+                                else:
+                                    yield data["t"]          
                            else:
-                                yield data["t"]                         
+                                yield data["t"]               

    @classmethod
    async def upload_file(cls, client: StreamSession, cookies: Cookies, file: bytes, filename: str = None) -> dict:
--- a/g4f/Provider/init.py
+++ b/g4f/Provider/init.py
@ -9,7 +9,6 @@ from .deprecated      import *
 from .not_working     import *
 from .selenium        import *
 from .needs_auth      import *
-from .unfinished      import *

 from .Aichatos         import Aichatos
 from .Aura             import Aura
@ -46,6 +45,7 @@ from .MetaAI           import MetaAI
 from .MetaAIAccount    import MetaAIAccount
 from .PerplexityLabs   import PerplexityLabs
 from .Pi               import Pi
+from .Replicate        import Replicate
 from .ReplicateImage   import ReplicateImage
 from .Vercel           import Vercel
 from .WhiteRabbitNeo   import WhiteRabbitNeo
--- a/g4f/Provider/bing/conversation.py
+++ b/g4f/Provider/bing/conversation.py
@ -41,6 +41,8 @@ async def create_conversation(session: StreamSession, headers: dict, tone: str)
            raise RateLimitError("Response 404: Do less requests and reuse conversations")
        await raise_for_status(response, "Failed to create conversation")
        data = await response.json()
+    if not data:
+        raise RuntimeError('Empty response: Failed to create conversation')
    conversationId = data.get('conversationId')
    clientId = data.get('clientId')
    conversationSignature = response.headers.get('X-Sydney-Encryptedconversationsignature')
--- a/g4f/Provider/needs_auth/Gemini.py
+++ b/g4f/Provider/needs_auth/Gemini.py
@ -16,6 +16,7 @@ try:
 except ImportError:
    pass

+from ... import debug
 from ...typing import Messages, Cookies, ImageType, AsyncResult
 from ..base_provider import AsyncGeneratorProvider
 from ..helper import format_prompt, get_cookies
@ -53,6 +54,56 @@ class Gemini(AsyncGeneratorProvider):
    url = "https://gemini.google.com"
    needs_auth = True
    working = True
+    image_models = ["gemini"]
+    default_vision_model = "gemini"
+    _cookies: Cookies = None
+
+    @classmethod
+    async def nodriver_login(cls) -> Cookies:
+        try:
+            import nodriver as uc
+        except ImportError:
+            return
+        try:
+            from platformdirs import user_config_dir
+            user_data_dir = user_config_dir("g4f-nodriver")
+        except:
+            user_data_dir = None
+        if debug.logging:
+            print(f"Open nodriver with user_dir: {user_data_dir}")
+        browser = await uc.start(user_data_dir=user_data_dir)
+        page = await browser.get(f"{cls.url}/app")
+        await page.select("div.ql-editor.textarea", 240)
+        cookies = {}
+        for c in await page.browser.cookies.get_all():
+            if c.domain.endswith(".google.com"):
+                cookies[c.name] = c.value
+        await page.close()
+        return cookies
+
+    @classmethod
+    async def webdriver_login(cls, proxy: str):
+        driver = None
+        try:
+            driver = get_browser(proxy=proxy)
+            try:
+                driver.get(f"{cls.url}/app")
+                WebDriverWait(driver, 5).until(
+                    EC.visibility_of_element_located((By.CSS_SELECTOR, "div.ql-editor.textarea"))
+                )
+            except:
+                login_url = os.environ.get("G4F_LOGIN_URL")
+                if login_url:
+                    yield f"Please login: [Google Gemini]({login_url})\n\n"
+                WebDriverWait(driver, 240).until(
+                    EC.visibility_of_element_located((By.CSS_SELECTOR, "div.ql-editor.textarea"))
+                )
+            cls._cookies = get_driver_cookies(driver)
+        except MissingRequirementsError:
+            pass
+        finally:
+            if driver:
+                driver.close()

    @classmethod
    async def create_async_generator(
@ -72,47 +123,30 @@ class Gemini(AsyncGeneratorProvider):
            if cookies is None:
                cookies = {}
            cookies["__Secure-1PSID"] = api_key
-        cookies = cookies if cookies else get_cookies(".google.com", False, True)
+        cls._cookies = cookies or cls._cookies or get_cookies(".google.com", False, True)
        base_connector = get_connector(connector, proxy)
        async with ClientSession(
            headers=REQUEST_HEADERS,
            connector=base_connector
        ) as session:
-            snlm0e  = await cls.fetch_snlm0e(session, cookies) if cookies else None
+            snlm0e  = await cls.fetch_snlm0e(session, cls._cookies) if cls._cookies else None
            if not snlm0e:
-                driver = None
-                try:
-                    driver = get_browser(proxy=proxy)
-                    try:
-                        driver.get(f"{cls.url}/app")
-                        WebDriverWait(driver, 5).until(
-                            EC.visibility_of_element_located((By.CSS_SELECTOR, "div.ql-editor.textarea"))
-                        )
-                    except:
-                        login_url = os.environ.get("G4F_LOGIN_URL")
-                        if login_url:
-                            yield f"Please login: [Google Gemini]({login_url})\n\n"
-                        WebDriverWait(driver, 240).until(
-                            EC.visibility_of_element_located((By.CSS_SELECTOR, "div.ql-editor.textarea"))
-                        )
-                    cookies = get_driver_cookies(driver)
-                except MissingRequirementsError:
-                    pass
-                finally:
-                    if driver:
-                        driver.close()
+                cls._cookies = await cls.nodriver_login();
+                if cls._cookies is None:
+                    async for chunk in cls.webdriver_login(proxy):
+                        yield chunk

            if not snlm0e:
-                if "__Secure-1PSID" not in cookies:
+                if "__Secure-1PSID" not in cls._cookies:
                    raise MissingAuthError('Missing "__Secure-1PSID" cookie')
-                snlm0e = await cls.fetch_snlm0e(session, cookies)
+                snlm0e = await cls.fetch_snlm0e(session, cls._cookies)
            if not snlm0e:
-                raise RuntimeError("Invalid auth. SNlM0e not found")
+                raise RuntimeError("Invalid cookies. SNlM0e not found")

            image_url = await cls.upload_image(base_connector, to_bytes(image), image_name) if image else None
        
            async with ClientSession(
-                cookies=cookies,
+                cookies=cls._cookies,
                headers=REQUEST_HEADERS,
                connector=base_connector,
            ) as client:
--- a/g4f/Provider/needs_auth/OpenaiAccount.py
+++ b/g4f/Provider/needs_auth/OpenaiAccount.py
@ -3,4 +3,6 @@ from __future__ import annotations
 from .OpenaiChat import OpenaiChat

 class OpenaiAccount(OpenaiChat):
-    needs_auth = True
+    needs_auth = True
+    parent = "OpenaiChat"
+    image_models = ["dall-e"]
--- a/g4f/Provider/needs_auth/OpenaiChat.py
+++ b/g4f/Provider/needs_auth/OpenaiChat.py
@ -29,6 +29,7 @@ from ...requests.aiohttp import StreamSession
 from ...image import to_image, to_bytes, ImageResponse, ImageRequest
 from ...errors import MissingAuthError, ResponseError
 from ...providers.conversation import BaseConversation
+from ..helper import format_cookies
 from ..openai.har_file import getArkoseAndAccessToken, NoValidHarFileError
 from ... import debug

@ -43,8 +44,14 @@ class OpenaiChat(AsyncGeneratorProvider, ProviderModelMixin):
    supports_message_history = True
    supports_system_message = True
    default_model = None
+    default_vision_model = "gpt-4-vision"
    models = ["gpt-3.5-turbo", "gpt-4", "gpt-4-gizmo"]
-    model_aliases = {"text-davinci-002-render-sha": "gpt-3.5-turbo", "": "gpt-3.5-turbo", "gpt-4-turbo-preview": "gpt-4"}
+    model_aliases = {
+        "text-davinci-002-render-sha": "gpt-3.5-turbo",
+        "": "gpt-3.5-turbo",
+        "gpt-4-turbo-preview": "gpt-4",
+        "dall-e": "gpt-4",
+    }
    _api_key: str = None
    _headers: dict = None
    _cookies: Cookies = None
@ -334,9 +341,8 @@ class OpenaiChat(AsyncGeneratorProvider, ProviderModelMixin):
        Raises:
            RuntimeError: If an error occurs during processing.
        """
-
        async with StreamSession(
-            proxies={"all": proxy},
+            proxy=proxy,
            impersonate="chrome",
            timeout=timeout
        ) as session:
@ -358,26 +364,27 @@ class OpenaiChat(AsyncGeneratorProvider, ProviderModelMixin):
                    api_key = cls._api_key = None
                    cls._create_request_args()
                    if debug.logging:
-                        print("OpenaiChat: Load default_model failed")
+                        print("OpenaiChat: Load default model failed")
                        print(f"{e.__class__.__name__}: {e}")

            arkose_token = None
            if cls.default_model is None:
+                error = None
                try:
-                    arkose_token, api_key, cookies = await getArkoseAndAccessToken(proxy)
-                    cls._create_request_args(cookies)
+                    arkose_token, api_key, cookies, headers = await getArkoseAndAccessToken(proxy)
+                    cls._create_request_args(cookies, headers)
                    cls._set_api_key(api_key)
                except NoValidHarFileError as e:
-                    ...
+                    error = e
                if cls._api_key is None:
                    await cls.nodriver_access_token()
                if cls._api_key is None and cls.needs_auth:
-                    raise e
+                    raise error
                cls.default_model = cls.get_model(await cls.get_default_model(session, cls._headers))

            async with session.post(
                f"{cls.url}/backend-anon/sentinel/chat-requirements"
-                if not cls._api_key else
+                if cls._api_key is None else
                f"{cls.url}/backend-api/sentinel/chat-requirements",
                json={"conversation_mode_kind": "primary_assistant"},
                headers=cls._headers
@ -393,8 +400,8 @@ class OpenaiChat(AsyncGeneratorProvider, ProviderModelMixin):
                    print(f'Arkose: {need_arkose} Turnstile: {data["turnstile"]["required"]}')

            if need_arkose and arkose_token is None:
-                arkose_token, api_key, cookies = await getArkoseAndAccessToken(proxy)
-                cls._create_request_args(cookies)
+                arkose_token, api_key, cookies, headers = await getArkoseAndAccessToken(proxy)
+                cls._create_request_args(cookies, headers)
                cls._set_api_key(api_key)
                if arkose_token is None:
                    raise MissingAuthError("No arkose token found in .har file")
@ -406,7 +413,8 @@ class OpenaiChat(AsyncGeneratorProvider, ProviderModelMixin):
                    print("OpenaiChat: Upload image failed")
                    print(f"{e.__class__.__name__}: {e}")

-            model = cls.get_model(model).replace("gpt-3.5-turbo", "text-davinci-002-render-sha")
+            model = cls.get_model(model)
+            model = "text-davinci-002-render-sha" if model == "gpt-3.5-turbo" else model
            if conversation is None:
                conversation = Conversation(conversation_id, str(uuid.uuid4()) if parent_id is None else parent_id)
            else:
@ -613,7 +621,7 @@ this.fetch = async (url, options) => {
                cookies[c.name] = c.value
        user_agent = await page.evaluate("window.navigator.userAgent")
        await page.close()
-        cls._create_request_args(cookies, user_agent)
+        cls._create_request_args(cookies, user_agent=user_agent)
        cls._set_api_key(api_key)

    @classmethod
@ -667,16 +675,12 @@ this.fetch = async (url, options) => {
            "oai-language": "en-US",
        }

-    @staticmethod
-    def _format_cookies(cookies: Cookies):
-        return "; ".join(f"{k}={v}" for k, v in cookies.items() if k != "access_token")
-
    @classmethod
-    def _create_request_args(cls, cookies: Cookies = None, user_agent: str = None):
-        cls._headers = cls.get_default_headers()
+    def _create_request_args(cls, cookies: Cookies = None, headers: dict = None, user_agent: str = None):
+        cls._headers = cls.get_default_headers() if headers is None else headers
        if user_agent is not None:
            cls._headers["user-agent"] = user_agent
-        cls._cookies = {} if cookies is None else cookies
+        cls._cookies = {} if cookies is None else {k: v for k, v in cookies.items() if k != "access_token"}
        cls._update_cookie_header()

    @classmethod
@ -693,7 +697,7 @@ this.fetch = async (url, options) => {

    @classmethod
    def _update_cookie_header(cls):
-        cls._headers["cookie"] = cls._format_cookies(cls._cookies)
+        cls._headers["cookie"] = format_cookies(cls._cookies)

 class Conversation(BaseConversation):
    """
--- a/g4f/Provider/openai/crypt.py
+++ b/g4f/Provider/openai/crypt.py
@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import json
 import base64
 import hashlib
--- a/g4f/Provider/openai/har_file.py
+++ b/g4f/Provider/openai/har_file.py
@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import base64
 import json
 import os
@ -59,17 +61,21 @@ def readHAR():
                    except KeyError:
                        continue
                    cookies = {c['name']: c['value'] for c in v['request']['cookies']}
+                    headers = get_headers(v)
    if not accessToken:
        raise NoValidHarFileError("No accessToken found in .har files")
    if not chatArks:
-        return None, accessToken, cookies
-    return chatArks.pop(), accessToken, cookies
+        return None, accessToken, cookies, headers
+    return chatArks.pop(), accessToken, cookies, headers
+
+def get_headers(entry) -> dict:
+    return {h['name'].lower(): h['value'] for h in entry['request']['headers'] if h['name'].lower() not in ['content-length', 'cookie'] and not h['name'].startswith(':')}

 def parseHAREntry(entry) -> arkReq:
    tmpArk = arkReq(
        arkURL=entry['request']['url'],
        arkBx="",
-        arkHeader={h['name'].lower(): h['value'] for h in entry['request']['headers'] if h['name'].lower() not in ['content-length', 'cookie'] and not h['name'].startswith(':')},
+        arkHeader=get_headers(entry),
        arkBody={p['name']: unquote(p['value']) for p in entry['request']['postData']['params'] if p['name'] not in ['rnd']},
        arkCookies={c['name']: c['value'] for c in entry['request']['cookies']},
        userAgent=""
@ -123,11 +129,11 @@ def getN() -> str:
    timestamp = str(int(time.time()))
    return base64.b64encode(timestamp.encode()).decode()

-async def getArkoseAndAccessToken(proxy: str):
+async def getArkoseAndAccessToken(proxy: str) -> tuple[str, str, dict, dict]:
    global chatArk, accessToken, cookies
    if chatArk is None or accessToken is None:
-        chatArk, accessToken, cookies = readHAR()
+        chatArk, accessToken, cookies, headers = readHAR()
    if chatArk is None:
-        return None, accessToken, cookies
+        return None, accessToken, cookies, headers
    newReq = genArkReq(chatArk)
-    return await sendRequest(newReq, proxy), accessToken, cookies
+    return await sendRequest(newReq, proxy), accessToken, cookies, headers
--- a/g4f/Provider/unfinished/Replicate.py
+++ b/g4f/Provider/unfinished/Replicate.py
@ -1,78 +0,0 @@
-from __future__ import annotations
-
-import asyncio
-
-from ..base_provider import AsyncGeneratorProvider, ProviderModelMixin
-from ..helper import format_prompt, filter_none
-from ...typing import AsyncResult, Messages
-from ...requests import StreamSession, raise_for_status
-from ...image import ImageResponse
-from ...errors import ResponseError, MissingAuthError
-
-class Replicate(AsyncGeneratorProvider, ProviderModelMixin):
-    url = "https://replicate.com"
-    working = True
-    default_model = "mistralai/mixtral-8x7b-instruct-v0.1"
-    api_base = "https://api.replicate.com/v1/models/"
-
-    @classmethod
-    async def create_async_generator(
-        cls,
-        model: str,
-        messages: Messages,
-        api_key: str = None,
-        proxy: str = None,
-        timeout: int = 180,
-        system_prompt: str = None,
-        max_new_tokens: int = None,
-        temperature: float = None,
-        top_p: float = None,
-        top_k: float = None,
-        stop: list = None,
-        extra_data: dict = {},
-        headers: dict = {},
-        **kwargs
-    ) -> AsyncResult:
-        model = cls.get_model(model)
-        if api_key is None:
-            raise MissingAuthError("api_key is missing")
-        headers["Authorization"] = f"Bearer {api_key}"
-        async with StreamSession(
-            proxies={"all": proxy},
-            headers=headers,
-            timeout=timeout
-        ) as session:
-            data = {
-                "stream": True,
-                "input": {
-                    "prompt": format_prompt(messages),
-                    **filter_none(
-                        system_prompt=system_prompt,
-                        max_new_tokens=max_new_tokens,
-                        temperature=temperature,
-                        top_p=top_p,
-                        top_k=top_k,
-                        stop_sequences=",".join(stop) if stop else None
-                    ),
-                    **extra_data
-                },
-            }
-            url = f"{cls.api_base.rstrip('/')}/{model}/predictions"
-            async with session.post(url, json=data) as response:
-                await raise_for_status(response)
-                result = await response.json()
-            if "id" not in result:
-                raise ResponseError(f"Invalid response: {result}")
-            async with session.get(result["urls"]["stream"], headers={"Accept": "text/event-stream"}) as response:
-                await raise_for_status(response)
-                event = None
-                async for line in response.iter_lines():
-                    if line.startswith(b"event: "):
-                        event = line[7:]
-                    elif event == b"output":
-                        if line.startswith(b"data: "):
-                            yield line[6:].decode()
-                        elif not line.startswith(b"id: "):
-                            continue#yield "+"+line.decode()
-                    elif event == b"done":
-                        break
--- a/g4f/Provider/you/har_file.py
+++ b/g4f/Provider/you/har_file.py
@ -4,7 +4,6 @@ import json
 import os
 import os.path
 import random
-import requests

 from ...requests import StreamSession, raise_for_status
 from ...errors import MissingRequirementsError
@ -21,7 +20,8 @@ class arkReq:
        self.arkCookies = arkCookies
        self.userAgent = userAgent

-arkPreURL = "https://telemetry.stytch.com/submit"
+telemetry_url = "https://telemetry.stytch.com/submit"
+public_token = "public-token-live-507a52ad-7e69-496b-aee0-1c9863c7c819"
 chatArks: list = None

 def readHAR():
@ -44,7 +44,7 @@ def readHAR():
                # Error: not a HAR file!
                continue
            for v in harFile['log']['entries']:
-                if arkPreURL in v['request']['url']:
+                if v['request']['url'] == telemetry_url:
                    chatArks.append(parseHAREntry(v))
    if not chatArks:
        raise NoValidHarFileError("No telemetry in .har files found")
@ -62,24 +62,29 @@ def parseHAREntry(entry) -> arkReq:
    return tmpArk

 async def sendRequest(tmpArk: arkReq, proxy: str = None):
-    async with StreamSession(headers=tmpArk.arkHeaders, cookies=tmpArk.arkCookies, proxies={"all": proxy}) as session:
+    async with StreamSession(headers=tmpArk.arkHeaders, cookies=tmpArk.arkCookies, proxy=proxy) as session:
        async with session.post(tmpArk.arkURL, data=tmpArk.arkBody) as response:
            await raise_for_status(response)
            return await response.text()

-async def get_dfp_telemetry_id(proxy: str = None):
+async def create_telemetry_id(proxy: str = None):
    global chatArks
    if chatArks is None:
        chatArks = readHAR()
    return await sendRequest(random.choice(chatArks), proxy)

 async def get_telemetry_ids(proxy: str = None) -> list:
+    try:
+        return [await create_telemetry_id(proxy)]
+    except NoValidHarFileError as e:
+        if debug.logging:
+            print(e)
    if debug.logging:
        print('Getting telemetry_id for you.com with nodriver')
    try:
        from nodriver import start
    except ImportError:
-        raise MissingRequirementsError('Install "nodriver" package | pip install -U nodriver')
+        raise MissingRequirementsError('Add .har file from you.com or install "nodriver" package | pip install -U nodriver')
    try:
        browser = await start()
        tab = browser.main_tab
@ -89,49 +94,11 @@ async def get_telemetry_ids(proxy: str = None) -> list:
            await tab.sleep(1)

        async def get_telemetry_id():
-            public_token = "public-token-live-507a52ad-7e69-496b-aee0-1c9863c7c819"
-            telemetry_url = "https://telemetry.stytch.com/submit"
-            return await tab.evaluate(f'this.GetTelemetryID("{public_token}", "{telemetry_url}");', await_promise=True)
+            return await tab.evaluate(
+                f'this.GetTelemetryID("{public_token}", "{telemetry_url}");',
+                await_promise=True
+            )

-        # for _ in range(500):
-        #     with open("hardir/you.com_telemetry_ids.txt", "a") as f:
-        #         f.write((await get_telemetry_id()) + "\n")
-
-        return [await get_telemetry_id() for _ in range(4)]
+        return [await get_telemetry_id() for _ in range(1)]
    finally:
-        try:
-            await tab.close()
-        except Exception as e:
-            print(f"Error occurred while closing tab: {str(e)}")
-        try:
-            await browser.stop()
-        except Exception as e:
-            pass
-
-    headers = {
-        'Accept': '*/*',
-        'Accept-Language': 'en,fr-FR;q=0.9,fr;q=0.8,es-ES;q=0.7,es;q=0.6,en-US;q=0.5,am;q=0.4,de;q=0.3',
-        'Connection': 'keep-alive',
-        'Content-type': 'application/x-www-form-urlencoded',
-        'Origin': 'https://you.com',
-        'Referer': 'https://you.com/',
-        'Sec-Fetch-Dest': 'empty',
-        'Sec-Fetch-Mode': 'cors',
-        'Sec-Fetch-Site': 'cross-site',
-        'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36',
-        'sec-ch-ua': '"Google Chrome";v="123", "Not:A-Brand";v="8", "Chromium";v="123"',
-        'sec-ch-ua-mobile': '?0',
-        'sec-ch-ua-platform': '"macOS"',
-    }
-    
-    proxies = {
-            'http': proxy,
-            'https': proxy} if proxy else None
-
-    response = requests.post('https://telemetry.stytch.com/submit', 
-                            headers=headers, data=payload, proxies=proxies)
-
-    if '-' in response.text:
-        print(f'telemetry generated: {response.text}')
-    
-    return (response.text)
+        await browser.stop()
--- a/g4f/api/init.py
+++ b/g4f/api/init.py
@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import logging
 import json
 import uvicorn
@ -8,14 +10,19 @@ from fastapi.exceptions import RequestValidationError
 from starlette.status import HTTP_422_UNPROCESSABLE_ENTITY
 from fastapi.encoders import jsonable_encoder
 from pydantic import BaseModel
-from typing import List, Union, Optional
+from typing import Union, Optional

 import g4f
 import g4f.debug
 from g4f.client import AsyncClient
 from g4f.typing import Messages

-app = FastAPI()
+def create_app() -> FastAPI:
+    app = FastAPI()
+    api = Api(app)
+    api.register_routes()
+    api.register_validation_exception_handler()
+    return app

 class ChatCompletionsConfig(BaseModel):
    messages: Messages
@ -29,16 +36,19 @@ class ChatCompletionsConfig(BaseModel):
    web_search: Optional[bool] = None
    proxy: Optional[str] = None

+list_ignored_providers: list[str] = None
+
+def set_list_ignored_providers(ignored: list[str]):
+    global list_ignored_providers
+    list_ignored_providers = ignored
+
 class Api:
-    def __init__(self, list_ignored_providers: List[str] = None) -> None:
-        self.list_ignored_providers = list_ignored_providers
+    def __init__(self, app: FastAPI) -> None:
+        self.app = app
        self.client = AsyncClient()
-    
-    def set_list_ignored_providers(self, list: list):
-        self.list_ignored_providers = list

    def register_validation_exception_handler(self):
-        @app.exception_handler(RequestValidationError)
+        @self.app.exception_handler(RequestValidationError)
        async def validation_exception_handler(request: Request, exc: RequestValidationError):
            details = exc.errors()
            modified_details = []
@ -54,17 +64,17 @@ class Api:
            )

    def register_routes(self):
-        @app.get("/")
+        @self.app.get("/")
        async def read_root():
            return RedirectResponse("/v1", 302)

-        @app.get("/v1")
+        @self.app.get("/v1")
        async def read_root_v1():
            return HTMLResponse('g4f API: Go to '
                                '<a href="/v1/chat/completions">chat/completions</a> '
                                'or <a href="/v1/models">models</a>.')

-        @app.get("/v1/models")
+        @self.app.get("/v1/models")
        async def models():
            model_list = dict(
                (model, g4f.models.ModelUtils.convert[model])
@ -78,7 +88,7 @@ class Api:
            } for model_id, model in model_list.items()]
            return JSONResponse(model_list)

-        @app.get("/v1/models/{model_name}")
+        @self.app.get("/v1/models/{model_name}")
        async def model_info(model_name: str):
            try:
                model_info = g4f.models.ModelUtils.convert[model_name]
@ -91,7 +101,7 @@ class Api:
            except:
                return JSONResponse({"error": "The model does not exist."})

-        @app.post("/v1/chat/completions")
+        @self.app.post("/v1/chat/completions")
        async def chat_completions(config: ChatCompletionsConfig, request: Request = None, provider: str = None):
            try:
                config.provider = provider if config.provider is None else config.provider
@ -103,7 +113,7 @@ class Api:
                            config.api_key = auth_header
                response = self.client.chat.completions.create(
                    **config.dict(exclude_none=True),
-                    ignored=self.list_ignored_providers
+                    ignored=list_ignored_providers
                )
            except Exception as e:
                logging.exception(e)
@ -125,14 +135,10 @@ class Api:

            return StreamingResponse(streaming(), media_type="text/event-stream")

-        @app.post("/v1/completions")
+        @self.app.post("/v1/completions")
        async def completions():
            return Response(content=json.dumps({'info': 'Not working yet.'}, indent=4), media_type="application/json")

-api = Api()
-api.register_routes()
-api.register_validation_exception_handler()
-
 def format_exception(e: Exception, config: ChatCompletionsConfig) -> str:
    last_provider = g4f.get_last_provider(True)
    return json.dumps({
@ -156,4 +162,4 @@ def run_api(
        host, port = bind.split(":")
    if debug:
        g4f.debug.logging = True
-    uvicorn.run("g4f.api:app", host=host, port=int(port), workers=workers, use_colors=use_colors)#
+    uvicorn.run("g4f.api:create_app", host=host, port=int(port), workers=workers, use_colors=use_colors, factory=True)#
--- a/g4f/cli.py
+++ b/g4f/cli.py
@ -1,7 +1,11 @@
+from __future__ import annotations
+
 import argparse

 from g4f import Provider
 from g4f.gui.run import gui_parser, run_gui_args
+from g4f.cookies import read_cookie_files
+from g4f import debug

 def main():
    parser = argparse.ArgumentParser(description="Run gpt4free")
@ -10,28 +14,36 @@ def main():
    api_parser.add_argument("--bind", default="0.0.0.0:1337", help="The bind string.")
    api_parser.add_argument("--debug", action="store_true", help="Enable verbose logging.")
    api_parser.add_argument("--workers", type=int, default=None, help="Number of workers.")
-    api_parser.add_argument("--disable_colors", action="store_true", help="Don't use colors.")
+    api_parser.add_argument("--disable-colors", action="store_true", help="Don't use colors.")
+    api_parser.add_argument("--ignore-cookie-files", action="store_true", help="Don't read .har and cookie files.")
    api_parser.add_argument("--ignored-providers", nargs="+", choices=[provider for provider in Provider.__map__],
                            default=[], help="List of providers to ignore when processing request.")
    subparsers.add_parser("gui", parents=[gui_parser()], add_help=False)

    args = parser.parse_args()
    if args.mode == "api":
-        import g4f.api
-        g4f.api.api.set_list_ignored_providers(
-            args.ignored_providers
-        )
-        g4f.api.run_api(
-            bind=args.bind,
-            debug=args.debug,
-            workers=args.workers,
-            use_colors=not args.disable_colors
-        )
+        run_api_args(args)
    elif args.mode == "gui":
        run_gui_args(args)
    else:
        parser.print_help()
        exit(1)

+def run_api_args(args):
+    if args.debug:
+        debug.logging = True
+    if not args.ignore_cookie_files:
+        read_cookie_files()
+    import g4f.api
+    g4f.api.set_list_ignored_providers(
+        args.ignored_providers
+    )
+    g4f.api.run_api(
+        bind=args.bind,
+        debug=args.debug,
+        workers=args.workers,
+        use_colors=not args.disable_colors
+    )
+
 if __name__ == "__main__":
    main()
--- a/g4f/cookies.py
+++ b/g4f/cookies.py
@ -2,6 +2,7 @@ from __future__ import annotations

 import os
 import time
+import json

 try:
    from platformdirs import user_config_dir
@ -38,6 +39,7 @@ def get_cookies(domain_name: str = '', raise_requirements_error: bool = True, si
    Returns:
        Dict[str, str]: A dictionary of cookie names and values.
    """
+    global _cookies
    if domain_name in _cookies:
        return _cookies[domain_name]

@ -46,6 +48,7 @@ def get_cookies(domain_name: str = '', raise_requirements_error: bool = True, si
    return cookies

 def set_cookies(domain_name: str, cookies: Cookies = None) -> None:
+    global _cookies
    if cookies:
        _cookies[domain_name] = cookies
    elif domain_name in _cookies:
@ -84,6 +87,61 @@ def load_cookies_from_browsers(domain_name: str, raise_requirements_error: bool
                print(f"Error reading cookies from {cookie_fn.__name__} for {domain_name}: {e}")
    return cookies

+def read_cookie_files(dirPath: str = "./har_and_cookies"):
+    global _cookies
+    harFiles = []
+    cookieFiles = []
+    for root, dirs, files in os.walk(dirPath):
+        for file in files:
+            if file.endswith(".har"):
+                harFiles.append(os.path.join(root, file))
+            elif file.endswith(".json"):
+                cookieFiles.append(os.path.join(root, file))
+    _cookies = {}
+    for path in harFiles:
+        with open(path, 'rb') as file:
+            try:
+                harFile = json.load(file)
+            except json.JSONDecodeError:
+                # Error: not a HAR file!
+                continue
+            if debug.logging:
+                print("Read .har file:", path)
+            new_cookies = {}
+            for v in harFile['log']['entries']:
+                v_cookies = {}
+                for c in v['request']['cookies']:
+                    if c['domain'] not in v_cookies:
+                        v_cookies[c['domain']] = {}
+                    v_cookies[c['domain']][c['name']] = c['value']
+                for domain, c in v_cookies.items():
+                    _cookies[domain] = c
+                    new_cookies[domain] = len(c)
+            if debug.logging:
+                for domain, new_values in new_cookies.items():
+                    print(f"Cookies added: {new_values} from {domain}")
+    for path in cookieFiles:
+        with open(path, 'rb') as file:
+            try:
+                cookieFile = json.load(file)
+            except json.JSONDecodeError:
+                # Error: not a json file!
+                continue
+            if not isinstance(cookieFile, list):
+                continue
+            if debug.logging:
+                print("Read cookie file:", path)
+            new_cookies = {}
+            for c in cookieFile:
+                if isinstance(c, dict) and "domain" in c:
+                    if c["domain"] not in new_cookies:
+                        new_cookies[c["domain"]] = {}
+                    new_cookies[c["domain"]][c["name"]] = c["value"]
+            for domain, new_values in new_cookies.items():
+                if debug.logging:
+                    print(f"Cookies added: {len(new_values)} from {domain}")
+                _cookies[domain] = new_values
+
 def _g4f(domain_name: str) -> list:
    """
    Load cookies from the 'g4f' browser (if exists).
--- a/g4f/gui/init.py
+++ b/g4f/gui/init.py
@ -12,9 +12,6 @@ def run_gui(host: str = '0.0.0.0', port: int = 8080, debug: bool = False) -> Non
    if import_error is not None:
        raise MissingRequirementsError(f'Install "gui" requirements | pip install -U g4f[gui]\n{import_error}')

-    if debug:
-        from g4f import debug
-        debug.logging = True
    config = {
        'host' : host,
        'port' : port,
--- a/g4f/gui/gui_parser.py
+++ b/g4f/gui/gui_parser.py
@ -5,4 +5,5 @@ def gui_parser():
    parser.add_argument("-host", type=str, default="0.0.0.0", help="hostname")
    parser.add_argument("-port", type=int, default=8080, help="port")
    parser.add_argument("-debug", action="store_true", help="debug mode")
+    parser.add_argument("--ignore-cookie-files", action="store_true", help="Don't read .har and cookie files.")
    return parser
--- a/g4f/gui/run.py
+++ b/g4f/gui/run.py
@ -1,6 +1,12 @@
 from .gui_parser import gui_parser
+from ..cookies import read_cookie_files
+import g4f.debug

 def run_gui_args(args):
+    if args.debug:
+        g4f.debug.logging = True
+    if not args.ignore_cookie_files:
+        read_cookie_files()
    from g4f.gui import run_gui
    host = args.host
    port = args.port
--- a/g4f/gui/server/api.py
+++ b/g4f/gui/server/api.py
@ -16,7 +16,8 @@ conversations: dict[dict[str, BaseConversation]] = {}

 class Api():

-    def get_models(self) -> list[str]:
+    @staticmethod
+    def get_models() -> list[str]:
        """
        Return a list of all models.

@ -27,7 +28,8 @@ class Api():
        """
        return models._all_models

-    def get_provider_models(self, provider: str) -> list[dict]:
+    @staticmethod
+    def get_provider_models(provider: str) -> list[dict]:
        if provider in __map__:
            provider: ProviderType = __map__[provider]
            if issubclass(provider, ProviderModelMixin):
@ -40,7 +42,28 @@ class Api():
            else:
                return [];

-    def get_providers(self) -> list[str]:
+    @staticmethod
+    def get_image_models() -> list[dict]:
+        image_models = []
+        for provider in __providers__:
+            if hasattr(provider, "image_models"):
+                if hasattr(provider, "get_models"):
+                    provider.get_models()
+                parent = provider
+                if hasattr(provider, "parent"):
+                    parent = __map__[provider.parent]
+                for model in provider.image_models:
+                    image_models.append({
+                        "provider": parent.__name__,
+                        "url": parent.url,
+                        "label": parent.label if hasattr(parent, "label") else None,
+                        "image_model": model,
+                        "vision_model": parent.default_vision_model if hasattr(parent, "default_vision_model") else None
+                    })
+        return image_models
+
+    @staticmethod
+    def get_providers() -> list[str]:
        """
        Return a list of all working providers.
        """
@ -58,7 +81,8 @@ class Api():
            if provider.working
        }

-    def get_version(self):
+    @staticmethod
+    def get_version():
        """
        Returns the current and latest version of the application.

--- a/g4f/gui/server/backend.py
+++ b/g4f/gui/server/backend.py
@ -31,6 +31,10 @@ class Backend_Api(Api):
                'function': self.get_provider_models,
                'methods': ['GET']
            },
+            '/backend-api/v2/image_models': {
+                'function': self.get_image_models,
+                'methods': ['GET']
+            },
            '/backend-api/v2/providers': {
                'function': self.get_providers,
                'methods': ['GET']
--- a/g4f/image.py
+++ b/g4f/image.py
@ -86,7 +86,7 @@ def is_data_uri_an_image(data_uri: str) -> bool:
    if image_format not in ALLOWED_EXTENSIONS and image_format != "svg+xml":
        raise ValueError("Invalid image format (from mime file type).")

-def is_accepted_format(binary_data: bytes) -> bool:
+def is_accepted_format(binary_data: bytes) -> str:
    """
    Checks if the given binary data represents an image with an accepted format.

@ -241,6 +241,13 @@ def to_bytes(image: ImageType) -> bytes:
    else:
        return image.read()

+def to_data_uri(image: ImageType) -> str:
+    if not isinstance(image, str):
+        data = to_bytes(image)
+        data_base64 = base64.b64encode(data).decode()
+        return f"data:{is_accepted_format(data)};base64,{data_base64}"
+    return image
+
 class ImageResponse:
    def __init__(
        self,
--- a/g4f/providers/base_provider.py
+++ b/g4f/providers/base_provider.py
@ -271,13 +271,13 @@ class AsyncGeneratorProvider(AsyncProvider):
        raise NotImplementedError()

 class ProviderModelMixin:
-    default_model: str
+    default_model: str = None
    models: list[str] = []
    model_aliases: dict[str, str] = {}

    @classmethod
    def get_models(cls) -> list[str]:
-        if not cls.models:
+        if not cls.models and cls.default_model is not None:
            return [cls.default_model]
        return cls.models

--- a/g4f/providers/helper.py
+++ b/g4f/providers/helper.py
@ -3,7 +3,7 @@ from __future__ import annotations
 import random
 import string

-from ..typing import Messages
+from ..typing import Messages, Cookies

 def format_prompt(messages: Messages, add_special_tokens=False) -> str:
    """
@ -56,4 +56,7 @@ def filter_none(**kwargs) -> dict:
        key: value
        for key, value in kwargs.items()
        if value is not None
-    }
+    }
+
+def format_cookies(cookies: Cookies) -> str:
+    return "; ".join([f"{k}={v}" for k, v in cookies.items()])
--- a/har_and_cookies/.gitkeep
+++ b/har_and_cookies/.gitkeep