Volcengine openviking is a context database for agents that treats memory, resources, and skills as…
volcengine openviking is a context database for agents that treats memory, resources, and skills as files in a directory instead of vectors in a store, and it just crossed 35,000 stars with the memory pipeline rebuilt in version 2. claude code, codex, cursor, and openclaw all ship plugins for it, and the work was accepted at vldb.
the next battleground is not model quality but what an agent remembers.
Context
The volcengine/OpenViking repository, read on 4 October 2026, shows 34,857 stars, a GNU AGPL-3.0 license and a creation date of 5 January 2026. The README describes an open-source context database for AI agents that stores memories, resources and skills as one virtual filesystem under a viking:// protocol, browsed with ls, tree and find. Its integrations list includes Claude Code, Codex, OpenClaw, Cursor and TRAE. It says the memory work is described in the VikingMem paper, arXiv 2605.29640, accepted by VLDB 2026, and that OpenViking open-sources a subset of what the paper describes.
The note says files instead of vectors, but the README says retrieval still uses vector search to find a directory first and then drills down. The 34,857 stars at fetch is below the note's 35,000 and neither proves nor refutes the figure on the day. The memory pipeline rebuilt in version 2 was not found in the README text read. The VLDB acceptance is the README's statement; the paper was not read. What an agent remembers being the next battleground is the author's opinion.
Watch next
- The arXiv paper and release notes.
Sources
- OpenViking (GitHub)github.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 18 September 2026 at 23:39 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →