Local two-agent setup shares one persistent memory and blocks bad decisions offline
PrajwalTomar_ · x · 2026-08-04
A builder describes a fully local setup where two AI agents share a persistent memory on a laptop—no cloud and no API keys.
- The “Coder” tried to import Postgres, even though an earlier decision had explicitly logged a rule to use SQLite instead.
- The “Reviewer” retrieved that old decision from shared memory and blocked the change by quoting the exact rule.
- After restarting the system, the same check still worked because the memory is stored on disk in Actian VectorAI DB.
- The larger point: most multi-agent systems forget decisions as soon as a session ends, and shared memory may be the missing piece that turns agents into a real team.
More from coding & agent
- ThursdAI explores whether Codex and multi-agent math can solve Erdős problems — thursdai_pod · 2026-08-04
- Digibee used Opik to version prompts and catch agent regressions before users did — dl_weekly · 2026-08-04
- Stripe built its internal Kai knowledge agent in one week with one engineer and Deep Agents — LangChain · 2026-08-04
- Using Claude with prebuilt shaders and assets to generate game worlds — mathemagic1an · 2026-08-04
- How I AI episode shows a workflow for handing voice, web and video tasks to Codex — nickbaumann_ · 2026-08-04
- Benchmark paper says some agent evals spend over $3,000 per task to elicit more capability — dejavucoder · 2026-08-04