Hey folks, I've been deep-diving into OpenPipe for the last few weeks, using it to fine-tune some smaller models on our internal chat logs. The platform itself is fantastic for the price, but I've hit a consistent snag: **data sync latency.**
I'm logging prompts/completions via their JavaScript SDK in a Next.js app, and I'm seeing delays of **several hours** before the logged data appears in the OpenPipe dashboard for training. This makes rapid iteration on prompt adjustments feel a bit like flying blind.
My setup is pretty straightforward:
```javascript
// app/api/chat/route.ts
import { OpenPipeClient } from "@openpipe/client";
const openpipe = new OpenPipeClient({
apiKey: process.env.OPENPIPE_API_KEY,
});
export async function POST(req: Request) {
// ... get messages
const startTime = Date.now();
const response = await openpipe.openAI({
model: "gpt-4",
messages,
});
const callLogId = response.openpipe?.callLogId;
// Later, report the final output
await openpipe.report({
callLogId,
completion: assistantMessage,
});
}
```
Is anyone else experiencing this? I've checked the obvious:
- API calls are returning `201` without errors.
- No quota limits are being hit.
- The data *does* eventually show up, just... later.
**My questions:**
1. Is this a known platform issue, or is it likely something on my end?
2. Are there any workarounds? (e.g., using the `cache` flag differently?)
3. Does the sync time improve on higher-tier plans, or is it just a processing queue thing?
I love the tool, but this lag makes the feedback loop for prompt tuning way longer than I'd hoped. Curious if others have run into this and how they've managed it.
--weaver