Developer postmortem: sync Anthropic client caused 11.4s voice latency
10/13/2026 — 10/13, 01:18·1 sources (all)·1 reports (all)
Key takeaways
- A developer published a postmortem of Preterview, a voice interview service, after latency rose from 0.9 seconds with one user to 11.4 secon…
Background & analysis
On October 13, 2026, a developer published a post on DEV Community reviewing a performance problem in Preterview, a voice interview service he had built. According to his account, with a single user on the server, the voice AI answered in 0.9 seconds. With eight users, the last person in line waited 11.4 seconds, and the service sounded broken. He says the model, the network, and the speech vendor were all fine. The issue came down to one line: client = Anthropic(), called from inside an async def. The synchronous Anthropic client was blocking the asyncio event loop used by FastAPI, so every other session on the same process stopped while that call ran. The developer says he spent two days blaming the wrong things before finding the cause.
AI-generated from 1 reports · updated 13 hours ago
Latest turnA developer recounts debugging his voice-interview app Preterview: answers took 0.9s with one user on the server, but the last person in line waited 11.4s with eight. The cause was a synchronous Anthropic client called inside an async def, which blocked the FastAPI asyncio event loop and stalled every session on the process. The write-up includes the numbers from the two-day hunt.
- Heat index
- 69
- All sources
- 1
- All reports
- 1
- First seen
- 13 hours ago
Reports on this story headlines open the original
A developer recounts debugging his voice-interview app Preterview: answers took 0.9s with one user on the server, but the last person in line waited 11.4s with eight. The cause was a synchronous Anthropic client called inside an async def, which blocked the FastAPI asyncio event loop and stalled every session on the process. The write-up includes the numbers from the two-day hunt.
DEV Community · AIAI score 60
Other stories people are talking about
- Heat index 59Anthropic Publishes Guide on Splitting Agent Instructions1 sources
- Heat index 159Anthropic commits $100M to train AI deployment engineers2 sources
- Heat index 470NewAnthropic Reportedly Considers Building Data Centers in Australia5 sources
- Heat index 277NewAnthropic New Security Measures Cut Model Distillation Data Theft by 90%+3 sources
- Heat index 1631RisingMeta launches Muse Wonder and Muse Wonder Full models21 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion