Vec1 version 0.7.
(1) By Dan Kennedy (dan) on 2026-07-07 14:29:35 [source]
Hi,
Vec1 is at version 0.7 now:
This is not a 1.0 release yet, but it's solid enough to stop calilng it "preview" or anything like that.
Latest version adds a "RabitQ" quantizer (vector compressor). This seems to work and has the advantage of not requiring training, but it doesn't really stack up to a proper trained quantizer like Product Quantization (PQ) or Optimized Product Quantization (OPQ) in terms of performance.
Vec1 should at this point be in-the-ballpark competitive with most ANN vector-search implementations for static datasets that can be trained on, or for data for which a representative sample is available to train on. It should also compare well to implementations of brute-force search.
Focus now shifts to adding a mode that does not require training at all - a mode that accepts slower writes in exchange for making them online and the index always available and searchable.
Any comments or questions are most welcome here!
Dan.
(2) By anonymous on 2026-07-27 07:12:25 in reply to 1 [link] [source]
Hello,
Just a tinkerer here on 10+-year-old hardware about to dip my toes into RAG with llama.cpp. Building and loading the extension was a breeze, will report back with test results!
Delighted to see SQLite-official version of this exists,
Chan
(3) By lsb (leebutterman) on 2026-08-02 16:58:14 in reply to 1 [link] [source]
Hi! I'm super excited to see this land at 1.0 and would love to help.
I'd seen in https://sqlite.org/vec1/doc/trunk/doc/vec1.md that in the Roadmap, "no further features are required" but "testing is insufficient". I've started to test vec1 with a coding agent, and have found with red-green TDD that there are numerous memory-safety bugs, including over half a dozen directly accessible from SQL.
Is this the right forum to report them? The AI-Agent Bug Forum? Privately?
Thanks,
Lee
(4) By Richard Hipp (drh) on 2026-08-02 17:39:31 in reply to 3 [link] [source]
For now, please report problems with Vec1 on the Vec1 Forum. Anonymous access has been activated as of a few moments ago. You can also now self-register on Vec1 at https://sqlite.org/vec1/login by pressing the "Create A New Account" button.
Thanks in advance for helping to test out Vec1.