# OpenAI for CLI completion?

**URL:** <https://community.openai.com/t/openai-for-cli-completion/10207>\
**Category:** API\
**Created:** [September 20, 2021, 3:53pm UTC](https://community.openai.com/t/openai-for-cli-completion/10207 "2021-09-20T15:53:54Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![louis030195](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/louis030195/32/36792_2.png) [@louis030195](https://community.openai.com/u/louis030195)\
**Post date:** [September 20, 2021, 3:53pm UTC](https://community.openai.com/t/openai-for-cli-completion/10207/1 "2021-09-20T15:53:54Z")

</div>

Is there such a thing online?

Would be nice to not have to memorize all the CLI stuff, especially for complex CLI such as `kubectl`or `gcloud`.

Example:  
I start typing `gcloud run --deploy terminator --` tab would propose, say, the `region` arg or `--max-instances` IDK

---

<div class="post-metadata">

**Author:** ![stevet](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/stevet/32/202_2.png) [@stevet](https://community.openai.com/u/stevet)\
**Post date:** [September 21, 2021, 1:21pm UTC](https://community.openai.com/t/openai-for-cli-completion/10207/2 "2021-09-21T13:21:17Z")

</div>

I’ve been using CoPilot and CodeX for doing something similar. I’m using them to create command line “cheatsheets” for the stuff I use from time-to-time. My guess is that it wouldn’t be too difficult to create a natural language CLI that would convert words into a CLI command. I really like your thinking here!

---

<div class="post-metadata">

**Author:** ![louis030195](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/louis030195/32/36792_2.png) [@louis030195](https://community.openai.com/u/louis030195)\
**Post date:** [September 22, 2021, 8:08am UTC](https://community.openai.com/t/openai-for-cli-completion/10207/3 "2021-09-22T08:08:29Z")

</div>

My idea was more about CLI completion rather than NLP CLI.

Just like we have tab completion currently on some CLI but more advanced, would guess the next part. When I started learning Kubernetes, I had to print a CLI cheat sheet because of its complexity… Humans are not good at memorization.

NLP CLI is dangerous IMHO, what if you say “un-deploy my backend-X12” and it misunderstands and does something like changing a backend version (which can lead, in my experience of days of downtime :D)

---

<div class="post-metadata">

**Author:** ![stevet](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/stevet/32/202_2.png) [@stevet](https://community.openai.com/u/stevet)\
**Post date:** [September 22, 2021, 9:37pm UTC](https://community.openai.com/t/openai-for-cli-completion/10207/4 "2021-09-22T21:37:38Z")

</div>

Good point for sure. You’d want to have some kind of validation step before submitting a command that was generated by NLP.