Agents lose context
Restarts break memory. Preferences, decisions, and prior work disappear unless memory is made portable.
Built on AMCP
Memory infrastructure
for AI agents
Nunchi AI builds connected products for thinking, building, and operating with long-term agent memory. The platform starts with AMCP, the open protocol for agent memory continuity, and extends into production tools.
AMCP
Nexus
Claude Code Channels
Protocol
AMCP
Portable memory continuity across sessions, runtimes, and clients.
Runtime
Nexus
Managed server for long-term agent memory and recall workflows.
Positioning
Small products, one tighter workflow.
Nunchi AI products stand alone, but they fit better together than apart.
Why this exists
Context breaks. Standards help. Products make it usable.
AMCP defines the layer
AMCP gives agent memory a transportable contract instead of leaving it trapped inside one product.
Nexus runs it in production
Nunchi AI turns the standard into a usable system with docs, hosted runtime, and connected workflows.
Latest proof
Validated through real workflows, not just protocol diagrams.
The public story is simple: AMCP is the memory layer, Nexus is the implementation, and Channels show where persistent context becomes immediately practical.
Open protocol
AMCP
Managed infrastructure
Nexus
Recent validation
Claude Code Channels
Writing
Notes, protocol drafts, and implementation thinking.
Claude Channels Makes the Agent Reachable. Nexus Makes It Remember.
A narrow verified proof that Claude Code Channels handles live reachability while Nexus preserves memory across restarts.
Read postWhy Agents Become Unstable as Conversations Get Longer
A practical explanation of why larger context windows do not automatically become reliable memory.
Read postCompleting the Agent Stack with Memory Continuity
Positioning AMCP as the missing memory layer beside tools and agent collaboration.
Read postProduct stack
A small ecosystem with clear roles.
Each product has a narrow job. Together they describe the operating surface of Nunchi AI.
Core platform
Nexus
Managed memory infrastructure for agents, built around AMCP and long-term recall workflows.
Open NexusCapture
Norfolk
Capture-first knowledge workflows that turn raw notes and sources into reusable memory.
Open NorfolkCompatibility
Vessel
Dependency and API change awareness before outdated assumptions hit production.
Open VesselOperations
Stackpilot
Developer operations automation packaged into a direct and narrow toolchain.
Open StackpilotMedia
Indie Pulse
Research and publishing workflows aimed at builders, products, and market signals.
Open Indie PulseCommerce
Qodr
Operational software for ordering flows and real-world service interfaces.
Open QodrKnowledge
Synapsis
Connected knowledge workflows where notes, structure, and agent memory meet.
Open SynapsisHow it fits
One platform narrative, multiple narrow products.
Protocol
AMCP defines how memory survives beyond one client or one session.
Infrastructure
Nexus exposes the memory system as something teams can actually run and adopt.
Applications
The surrounding products turn that memory layer into practical workflows across coding, media, and operations.