Public documentation for MoteCloud: memory infrastructure for agents and AI systems.
MoteCloud helps agents remember durable facts, prepare relevant context, stream working-session observations, and keep public, tenant, and operator knowledge in the right places. This repository contains the public documentation surface only.
- Overview - what MoteCloud is and where it fits.
- Quickstart - the first public path for getting oriented.
- Memory primitives - the public vocabulary for memory items, metadata, sessions, and relationships.
- Context priming - how agents receive useful memory context.
- Agent memory - the recommended lifecycle for agent-facing memory.
- Retrieval and graphs - how MoteCloud combines relevance and relationships at a product level.
- Privacy and documentation boundaries - what belongs in public docs, tenant docs, and admin docs.
- Using MoteCloud with MCP - the public MCP integration model.
- Using the API - public guidance for API integrations.
- Using agents - practical patterns for agent workflows.
- Benchmark overview - how public benchmark material is organized.
- Tenant onboarding
- Publishing benchmarks
- API reference overview
- MCP tools reference
- Environment reference
- Error guide
- Public status guide
- Support guide
This repository is intentionally curated. It can include product explanations, public quickstarts, API and MCP usage guides, benchmark methodology, changelog entries, and support information.
It does not include private source links, tenant-specific data, admin runbooks, security evidence, internal strategy, unreleased research, credentials, deployment secrets, or private benchmark artifacts.
- Website: motecloud.io
- Public benchmarks: github.com/motecloud/benchmarks
- Changelog: CHANGELOG.md
- Security reporting: SECURITY.md
This repository starts with curated public guides. Reference pages and generated specifications should be added only after they are reviewed for public-safety boundaries.