January 13, 2026
Killed it, it lived
Show HN: Ayder – HTTP-native durable event log written in C (curl as client)
Dev ships crash-proof 'message inbox' you curl; fans rave, skeptics yell AI
TLDR: Ayder is a lightweight, crash-resilient message system you can use with curl, promising strong durability without heavy setup. The community is split between loving its simplicity and speed, suspecting the README’s swagger, and debating Raft vs newer tech like VSR—making it both intriguing and polarizing.
A new open-source tool, Ayder, just crash-landed into Hacker News—literally. The creator kicked off the demo by sending a kill signal to his own server to show it bounces back, and the crowd gasped, then clapped. Ayder promises a simple “message inbox” for apps—no heavy Java setups, just HTTP and curl—but with strong durability thanks to Raft (a way to keep multiple servers in agreement). Think: the middle ground between heavyweight Kafka and lightweight Redis Streams.
The vibes? Split. One camp is swooning over the one-binary, quick-start setup and millisecond-level delays. “This reminds me of NSQ,” said fans, resurrecting the cult classic and demanding it finally get its flowers. Another camp rolled in with skeptical side-eye, poking at the README’s swagger lines like “No ISR drama” (ISR means in-sync replicas) and asking if the doc was AI-generated. Meanwhile, the nerds dove into the deep end: “But what about TigerBeetle’s VSR?”—a more exotic consensus system—sparking a friendly protocol debate.
Between crash-test theatrics, 50k messages per second claims, and a Docker compose that brings dashboards in minutes, the thread turned into a mix of hype, homework, and heckling. Curious? Grab curl and try it yourself: GitHub repo
Key Points
- •Ayder is an HTTP-native, single-binary event streaming system written in C that uses curl as the client.
- •It aims for Kafka-grade durability via Raft sync-majority replication while maintaining Redis-like operational simplicity.
- •Features include append-only partitioned logs, consumer groups with committed offsets, sealed AOF with crash recovery, HA replication, a KV store, and streaming queries.
- •Benchmarks on a 3-node Raft cluster show 50,000 messages/second and P99 latency of ~3.35 ms; writes are replicated to a majority before acknowledgment.
- •Quick setup is available via Docker Compose with Prometheus and Grafana; source builds require libuv, OpenSSL, zlib, and liburing, with HTTP endpoints for produce/consume and offset commits.