OpenViking: open-source context database cuts agent input tokens by up to 91%
blaizedsouza · x · 2026-09-14
OpenViking is an open-source context database giving AI agents a single place to store knowledge, user memories, and reusable experience, organized as a virtual filesystem agents browse with familiar commands like ls, tree, and read. Instead of loading everything into context, summaries let the agent decide which files to open in full. Benchmarks report 80-83% memory accuracy with input tokens reduced by 34.3% to 91%.
Related event: OpenViking: Open-Source Context Database for AI Agents(2 posts)→
More from coding & agent
- Non-coders using AI: force agents to spell out design decisions and tradeoffs — brandon_galang · 2026-09-14
- One-sentence prompt fixes for Claude Fable 5.1's file-rewriting and token-burn habits — alex_verem · 2026-09-14
- Escape hatch guide: replace ChatGPT/Claude coding with OpenRouter + OpenCode for $0.61 — NewYak4281 · 2026-09-14
- LLM translation plus per-language glossaries: a localization workflow that works — Josikinz · 2026-09-14
- memx: an open-source CLI that debugs agent memory failures via state snapshot diffs — vivek9patel · 2026-09-14
- Application Agent Manifest proposal: ship desktop apps with an AI operator manual for agents — dzoricic · 2026-09-14