Reviewing Xiaohongshu's Enterprise AI Assistant Deployment
小红书技术REDtech · wechat · 2026-07-14
Xiaohongshu architects shared the deployment journey of their enterprise AI personal assistant, achieving full staff coverage in about a month: the decision was made in February, a beta was built by 3 people in 3 days, architecture was reinforced and public beta launched the following week, and company-wide rollout was completed within two weeks. The article details several core capabilities: solving security issues via an independent K8S cluster, MicroVM sandbox, and auditing mechanisms; reducing context token consumption via Self-GC; driving down costs using SealRouter for "on-demand model selection"; and managing short-term context, fact buffers, and long-term knowledge separately through a three-tier memory system. The author also introduced internal ecosystems like SkillHub and CoworkStudio, outlining a future shift from passive responses to proactive perception, goal-driven actions, and multi-Agent collaboration.
More from coding & agent
- Autoresearch proposes packaging ML runs as studies with questions, analysis, and code diffs — morgymcg · 2026-07-21
- CHAP defines approvals, handoffs, and audit logs for human-agent workflows — DeliveryTechnical199 · 2026-07-21
- The author says Codex reached 20x and is now debugging spec decoding on a hybrid parallel setup — TheZachMueller · 2026-07-21
- Axcess adds an MCP connector for WCAG accessibility checks that scanners miss — modelcontextprotocol · 2026-07-21
- X post asks whether Cursor Composer, built on Kimi models, would also be banned — max_paperclips · 2026-07-21
- A developer’s Codex usage is draining pooled enterprise credits at a small company — Distinct_Relation_62 · 2026-07-21