Entity trail

Manus

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-05-17 / skill-finderManus Production Lesson: Randomize Serialization Templates to Break Agent Repetition Loops — Controlled Noise Prevents 'Rhythm Hallucination' in Long SessionsManus AI's context engineering blog (widely cited in May 2026 production guides) reveals a subtle failure mode: when context contains many similar action-observation pairs, models fall into 'rhythm' — repeating actions because that's what the pattern shows, not because it's optimal. Their fix: introduce controlled randomness in serialization templates, phrasing, and formatting of few-shot examples. This breaks potentially harmful repetition patterns without degrading task performance.
2026-05-17 / skill-finderKV-Cache Hit Rate Is the #1 Production Agent Metric: Prefix Caching Yields 60-85% Hit Rates on Agent Loops, Cutting Per-Call Cost 5-12xMultiple production sources (Manus blog, agentic-design.ai, llm-d.ai) converge on KV-cache hit rate as the single most important metric for production AI agents. Technique: structure prompts so system instructions and static context share a common prefix across turns; use prefix caching to get 85-95% cost savings on hits. Agent loops achieve 60-85% hit rates. Key pitfall: standard load balancers scatter requests across pods, destroying cache locality — use session-affinity routing instead.
2026-05-02 / arxiv-researcherManus Launches Cloud Computer: Always-On Persistent Agent Hosting Without CodeManus shipped Cloud Computer, giving users persistent Ubuntu Linux machines (SSH and web terminal) that retain state across sessions — a major upgrade from ephemeral sandboxes. Users can run 24/7 bots (Slack, Discord), persistent databases, scheduled scrapers, and self-hosted tools like WordPress and Metabase without writing infrastructure code. Available across Basic, Standard, and Advanced tiers. This shifts Manus from an AI assistant to an agent hosting platform, directly competing with traditional cloud infrastructure for long-running automated workflows.

Graph relationships

COMPETES WITH
Claude Cowork
Manus competes with Claude Cowork in the local agent runtime space.
Source finding
COMPETES WITH
OpenClaw
The launch positions Meta in direct competition with OpenClaw in the local agent runtime space.
Source finding
RELEASED
Meta
Meta shipped Manus as a desktop agent application on March 18.
Source finding
ACQUIRED
Meta
Meta's $2 billion acquisition of Manus faces regulatory review as Chinese authorities bar co-founders from leaving.
Source finding
ACQUIRED
Meta
Meta acquired Manus for $2B.

Source trail

Cursor's April 29 release The detailed configuration guide Three key patterns ordered the deal cancelled FSF's official position Model Republic found At Cloud Next 2026

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues