# Has anyone created a CLI for ChatGPT-Plus Web Interface?

**URL:** <https://community.openai.com/t/has-anyone-created-a-cli-for-chatgpt-plus-web-interface/597591>\
**Category:** Feature requests\
**Tags:** gpt-4, chatgpt\
**Created:** [January 22, 2024, 11:06pm UTC](https://community.openai.com/t/has-anyone-created-a-cli-for-chatgpt-plus-web-interface/597591 "2024-01-22T23:06:45Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![synapticsage](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/synapticsage/32/34257_2.png) [@synapticsage](https://community.openai.com/u/synapticsage)\
**Post date:** [January 22, 2024, 11:06pm UTC](https://community.openai.com/t/has-anyone-created-a-cli-for-chatgpt-plus-web-interface/597591/1 "2024-01-22T23:06:45Z")

</div>

PLEASE NOTE: I AM AWARE OF THE OPENAI (GPT4) API AND USE IT… this is a different issue.

Yes, we there exists an API (separate from chatgpt-plus) and wonderful CLIs like charm/mods that exploit that API.

But I also have a wealth of less-expensive conversations in chatgpt-plus’s web interface that I would like to interact with at the command line.

Has anyone created a TUI-CLI or CLI for interacting with these conversations? E.g. grouping them into folders, bulk deleting conversations, bulk searching conversations?

---

<div class="post-metadata">

**Author:** ![Michael\_VAISSIERE](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/michael_vaissiere/32/646721_2.png) [@Michael\_VAISSIERE](https://community.openai.com/u/Michael_VAISSIERE)\
**Post date:** [June 25, 2025, 11:16am UTC](https://community.openai.com/t/has-anyone-created-a-cli-for-chatgpt-plus-web-interface/597591/2 "2025-06-25T11:16:03Z")

</div>

Hi!  
he **web version of ChatGPT Plus** is _so much more_ than just API access to a model. It’s a “living” memory, a knowledge space:

- **Rich, searchable conversation history** (with actual context and thread structure)
- **Folders and workspaces**
- **Projects** : contextual glue for long-term reasoning
- **Custom Instructions** : my own voice, my rules, my persona
- **Plugins, code interpreter, connectors**
- **Fine-tuned, evolving UI with context awareness**
- 

**So I’m left doing the only thing possible:**  
I have to _rebuild_ all of that, by hand, in Python.  
That means:

- Manual management of all conversation states, memory, and context windows
- Faking “projects” and “folders” with my own JSON or database structures
- Hard-coding my “custom instructions” in every single prompt
- Rolling my own embeddings and semantic search just to _find_ a conversation
- No way to bulk search, group, or organize web-based conversations from the CLI
- No easy way to “continue” a rich ChatGPT Plus thread outside the browser

**Let’s be honest:**  
It’s like sitting on the edge of a beautiful garden, peering through the gate, while only being allowed to plant seeds in a sandbox next door.

I know all the technical and security reasons why things are this way — but imagine the _potential_ if the “real” ChatGPT experience was available natively, with context, instructions, and all my creative work, right from the terminal. Imagine the workflows, the automations, the cross-platform magic.

**This is why I’m forced to reinvent the wheel, piecing together memory, embeddings, folders, custom instructions, and context management — all manually. And I know I’m not the only one.**

If anyone at OpenAI is listening:  
_Give us the keys to our own garden. Or at least, give us a well-lit footpath between worlds._