ByteDance open-sources OpenViking: agent memory that cuts input tokens up to 91%
Promptmethus · x · 2026-11-05
ByteDance has open-sourced OpenViking, a memory system for AI agents that organizes memory, knowledge, and skills like a file system: the agent reads short summaries first, then opens only the folders it needs.
- Long-chat memory tests show input tokens down 91% on OpenClaw, 63% on Claude Code, 34% on Hermes
- Accuracy rises from 24-57% to 80-83%
- One installer adds the memory plugin to Claude Code, Codex, Cursor, OpenCode and more; a server or their cloud setup works, plus a browser-based Studio
The original post estimates a user spending $11k/month on input tokens could save $10k. Responders note the concept echoes Jim Fan's Minecraft Voyager from the early agent era, but the productionized open-source implementation is drawing attention.
More from coding & agent
- Matt Pocock shares /writing-for-agents skill for writing content for AI agents — mattpocockuk · 2026-11-05
- Replace LLM judges with decision models across four RAG pipeline stages for speed and cost — kalyan_kpl · 2026-11-05
- oxus: MCP server lets agents search Mac via Spotlight instead of crawling with find — Able_Bluebird3654 · 2026-11-05
- Matt Pocock's one-line prompt makes coding agents verify their own changes — mattpocockuk · 2026-11-05
- Vercel's agent-browser CLI for AI agents hits 45.6k GitHub stars — vercel-labs · 2026-11-05
- Cloudflare manages AI spend with tiered alerts, not token limits, says CTO — vincent_koc · 2026-11-05