Shared my config for a social media post scheduler. Uses GPT-4, costs about $2/post.

Just finished a deep dive into SuperAGI for automating social media content, and I have to say, I'm really impressed with the flexibility. But as always, my brain immediately goes to the cost breakdown. 😅

I built an agent workflow that plans, writes, and formats a week's worth of posts for my tech blog's LinkedIn/Twitter. The key cost driver is using GPT-4 for the actual content generation. Here's a rough snapshot of my config and the math:

**Key Config Points:**
* **LLM:** GPT-4 (for high-quality output)
* **Tools:** Web scraper (for research), image generator (for simple graphics)
* **Workflow:** Research -> Outline -> Draft -> Review -> Format -> Schedule

**The Cost Breakdown:**
My typical post generation uses about 5,000 input and 1,000 output tokens with GPT-4. At current pricing, that's roughly:
* ~$0.15 for prompt processing
* ~$0.06 for the completion
* **Total per post:** ~$0.21 for the LLM call.

Add in a bit for the SuperAGI overhead, tool executions, and image generation, and I'm landing right around **$2 per complete, scheduled post**. That includes a custom graphic and platform-specific formatting.

For me, this is a win. Manual creation for a quality post took me at least an hour. Now I'm trading a few dollars for 10 minutes of review and tweaking. The ROI on time is huge.

Has anyone else built a similar content workflow? I'm curious if you've found ways to optimize the token usage further or switched models for certain steps to trim costs without losing quality.

— Jackson