kero
banner
kerollmops.com
kero
@kerollmops.com
Co-founder and Tech Lead at @meilisearch.com. You can find my blog on https://blog.kerollmops.com. I love #rustlang.
I think you should put the BufReader around the File and not the decoder 👀
October 30, 2025 at 8:06 PM
Reposted by kero
Learn more about Meilisearch Chat on our blog: www.meilisearch.com/blog/meilise...
Introducing Meilisearch Chat
Conversational search from your existing indexes with a single, OpenAI-compatible /chat endpoint. Built-in RAG, faster launches, lower costs.
www.meilisearch.com
October 13, 2025 at 9:17 AM
Reposted by kero
Without having numbers (right now), LMDB will probably read absurdly fast, while not so much space- or write efficient.
Most, if not all data, fits into RAM here. Would be interesting to benchmark 100s of GBs.

But I found redb to be slower (sometimes much slower) than LMDB, pretty much always.
October 11, 2025 at 2:22 PM
Have you figured out how heed/LMDB behaves on similar workloads? Because ReDB is highly inspired by it 👀
October 11, 2025 at 2:16 PM
Have you got the results then ? 🤓
October 5, 2025 at 4:06 PM
Ho crap! That's unfortunate, because LMDB is often in the top performers
October 2, 2025 at 5:42 PM
Have you ran these benchmarks for heed/LMDB too? 👀
October 2, 2025 at 5:38 PM
Ho no! I don't know if there is a recording of the talks 👀
September 25, 2025 at 12:52 PM
Would you mind defining "way more complex", please? Is the complexity in terms or debugging, publishing, maintaining, sharing, accepting external contributions?
September 20, 2025 at 10:19 AM