For client-facing workflows, where do local LLMs actually hold up vs cloud models?

Posted by Comfortable-Week7646@reddit | LocalLLaMA | View on Reddit | 3 comments

I’ve been going through older threads here, but most of what I found leans more toward benchmarks and model comparisons than actual day-to-day usage, so I figured I’d ask this from a practical angle.

For those using local LLMs in real client work (freelancing, agency stuff, etc.), where have they actually been reliable for you?

I’ve been experimenting with a mix of local and cloud for things like drafting proposals, summarizing client briefs, and organizing notes. Local models are great from a privacy and cost standpoint, but I still run into moments where the output just isn’t consistent enough depending on the task.

Right now everything feels a bit fragmented some workflows local, others still relying on cloud, and switching between tools more than I’d like.

Curious how others are handling this:

Would be really helpful to hear what’s actually working in practice.