PromptLayer alternatives that are not open-source self-hosted

Hey everyone, I've been using PromptLayer for a few months now to track my OpenAI and Anthropic calls. The dashboard and prompt versioning are super handy for my benchmark posts. However, I'm hitting some scaling cost concerns, and their team-based features feel a bit heavy for my solo experimentation.

I'm looking for alternatives, but with a specific constraint: **I do NOT want to self-host an open-source solution.** I've tried a few (like the LangSmith open-source beta) and the infra overhead just kills my vibe. I want a managed service.

My core needs:
* **Multi-provider logging** (OpenAI, Anthropic, maybe Gemini/Mistral)
* **Simple, searchable request tracing** to compare outputs.
* **Basic cost tracking** per project/model.
* An API/SDK that's as simple as PromptLayer's wrapper.

What I don't need (right now):
* Full-blown AI agent orchestration
* Extensive team RBAC
* Built-in eval suites (I roll my own)

I've glanced at:
* **LangSmith** (managed): Seems powerful but maybe overkill? The pricing is opaque.
* **Arize Phoenix**: More focused on evals and tracing, but can it do the simple logging?
* **Portkey**: Prompts are secondary; they seem more about gateway/load balancing.

Has anyone moved from PromptLayer to another *managed* service and been happy? Or found a lighter-weight hosted option? I'd love to see some code snippets of how the logging compares.

For example, here's my typical PromptLayer pattern:
```python
import promptlayer
openai = promptlayer.openai
response = openai.ChatCompletion.create(
model="gpt-4",
messages=[{"role": "user", "content": "Explain quantum tunneling."}],
pl_tags=["benchmark-round-3"]
)
```

What does the equivalent look like in your alternative?

--experiment