# Trying to learn

**URL:** <https://community.crewai.com/t/trying-to-learn/6490>\
**Category:** CrewAI Community Support\
**Tags:** agent\
**Created:** [July 6, 2025, 1:26pm UTC](https://community.crewai.com/t/trying-to-learn/6490 "2025-07-06T13:26:49Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![T\_P](https://sea1.discourse-cdn.com/flex025/user_avatar/community.crewai.com/t_p/32/1771_2.png) [@T\_P](https://community.crewai.com/u/T_P)\
**Post date:** [July 6, 2025, 1:26pm UTC](https://community.crewai.com/t/trying-to-learn/6490/1 "2025-07-06T13:26:49Z")

</div>

Hello,

First, I am incredibly new to any and all things related to AI,but very interested in learning and trying my best to do so. I am a little stumped at the moment and hoping anyone may be able to help me, I am sure its actually a simple fix.

I created my own agents in CrewStudio, but getting them to actually run and generate an output has stumped me. The crew must search the web, so for that to happen do I need to generate API keys from Serper Dev and have those input to the Enivormental Variables section? Additionally, I assume I need to do the same for the connection to GPT?

My next problem, I am receiving a message saying I have hit my limit on GPT. Is that simply a fix of buying more tokens? I haven’t actually been able to run my agents so I do not fully understand how I would have hit any limit, but again, I am not sure.

Any guidance would be incredibly appreciated.

---

<div class="post-metadata">

**Author:** ![Tony\_Wood](https://sea1.discourse-cdn.com/flex025/user_avatar/community.crewai.com/tony_wood/32/1876_2.png) [@Tony\_Wood](https://community.crewai.com/u/Tony_Wood)\
**Post date:** [July 7, 2025, 5:04pm UTC](https://community.crewai.com/t/trying-to-learn/6490/2 "2025-07-07T17:04:08Z")

</div>

Welcome to the community

A great place to start is [Installation - CrewAI](https://docs.crewai.com/en/installation)

For the limit. If you hit your limit try adding more tokens to see if it fixes it.

good luck on your journey

---

<div class="post-metadata">

**Author:** ![Harikrishnan\_K](https://sea1.discourse-cdn.com/flex025/user_avatar/community.crewai.com/harikrishnan_k/32/3848_2.png) [@Harikrishnan\_K](https://community.crewai.com/u/Harikrishnan_K)\
**Post date:** [July 11, 2025, 6:38am UTC](https://community.crewai.com/t/trying-to-learn/6490/3 "2025-07-11T06:38:16Z")

</div>

Since you’re just starting out, and if cost is a concern, I’d recommend trying **Ollama** for local model inference. It’s free and runs models like LLaMA or Mistral on your machine. It’s slower and less accurate than OpenAI GPT models, but good enough for learning . It won’t burn your API quota. Depending upon the complexity of tasks you are performing,system specification time consuming can differ.