v0.0.1

Release Notes for v0.0.1

This is the initial release of Memanto - a production-ready semantic long-term memory system for AI agents. Built on Moorcheh.ai’s no-indexing semantic database, Memanto delivers zero-cost ingestion latency, state-of-the-art accuracy (89.8% on LongMemEval, 87.1% on LoCoMo), instant write-to-retrieval, and true semantic search across conversations, sessions, and workflows.

Improvements

Semantic Memory Engine

• Agents: Persistent identity with isolated memory namespaces (e.g. customer-support-bot or dev-assistant).
• Sessions: 6-hour active windows; memories persist forever and remain accessible across all future sessions.
• Memories: Semantic units supporting 13 structured memory types (fact, preference, decision, goal, instruction, event, etc.) with confidence scoring.
• Zero-Indexing Semantic Search: Memories are immediately available for natural-language recall with no indexing delay or background processing.
• High-Accuracy Retrieval: Matches intent and context (e.g., “User prefers email communication” → “How should we contact the user?”).

Memanto CLI

• Complete memanto command-line interface installed via pip install memanto.
• Organized command groups: agent (create/activate/deactivate), memory (remember/recall/export), session, schedule, config, connect (integrations), and core utilities.
• Global options: --help, --version.
• Quickstart workflow:

  memanto                  # initial API key configuration
  memanto agent create my-agent
  memanto agent activate my-agent
  memanto remember "Project kickoff is Monday" --type event
  memanto recall "When is project kickoff?"

REST API & Authentication

• Full v2 HTTP API for programmatic memory management, agent lifecycle, recall, and generative answers.
• Get your Moorcheh API key from console
• Secure dual authentication:
  • Authorization: Bearer <moorcheh-api-key> for all requests.
  • X-Session-Token: <jwt> (6-hour session token obtained via /agents/{agent_id}/activate) for memory operations.
• JWT token introspection support and session extension endpoint included.
• Python httpx examples provided for both API key and session token flows.

Developer Integrations

• Native integrations with 13+ AI coding assistants and IDEs:
  • Claude Code, Cursor, Cline, Windsurf, Continue, GitHub Copilot, OpenCode, Goose, Roo, Antigravity, Augment, Gemini CLI, Codex.
• Simple connection via memanto connect <tool> (project-local or --global scope).
• Enables persistent context storage (preferences, decisions, architecture choices) and cross-session recall inside coding tools.
• Supports multi-tool and multi-environment setups (separate agents per tool or environment).

MemantoClaw

• Open-source reference stack for secure, always-on memory-augmented agents.
• Combines OpenClaw (autonomous agent framework), NVIDIA OpenShell (hardened sandbox with seccomp, Landlock, filesystem restrictions), and Memanto/Moorcheh memory.
• One-command provisioning: memantoclaw onboard (automatically configures inference routing, credentials, and zero-config Memanto memory bridge).
• Enhanced security: stricter policies than community OpenShell, credential filtering, immutable gateway config, and host-bridge memory architecture.

Getting Started

See the official quickstart in the Introduction and CLI Overview.
Deployment options include Docker, Python, AWS ECS, Google Cloud Run, Azure Container Instances, or local single-machine setups.

Documentation

• Full documentation: https://docs.memanto.ai
• CLI command reference, API authentication guide, integration setup, and MemantoClaw overview are all live.
• Moorcheh.ai base documentation: https://docs.moorcheh.ai

This v0.0.1 release establishes the complete Memanto ecosystem for production-grade long-term memory in AI agents and developer workflows.