memora
Memora is a lightweight MCP-compatible memory server that provides persistent semantic memory storage (SQLite with optional cloud sync), knowledge-graph linking, and a graph UI. It supports embeddings-based semantic search, RAG chat with tool calling, and LLM-assisted deduplication/merge workflows.
Score Breakdown
⚙ Agent Friendliness
🔒 Security
The README documents use of external provider API tokens/keys via environment variables (CLOUDFLARE_API_TOKEN, OPENAI_API_KEY, AWS credentials) and an option to encrypt cloud-uploaded DBs (MEMORA_CLOUD_ENCRYPT). However, there is no described authentication model for the MCP server itself, no explicit mention of TLS requirements for the local HTTP transport, no described authorization boundaries/scopes for tool actions, and no documented data-leak controls beyond storage encryption/compression.
⚡ Reliability
Best When
You want an MCP server-based memory layer that an agent can call locally (stdio MCP) or via a streaming HTTP transport, with optional cloud backing and a visualization UI.
Avoid When
You need strict, standardized API contracts (OpenAPI/SDK), or you cannot tolerate operational complexity around LLM providers, embeddings models, or cloud database credentials.
Use Cases
- • Persistent cross-session memory for AI agents
- • Semantic search over long-term notes and conversations
- • Building and visualizing knowledge graphs of memories and relationships
- • RAG chat over a personal knowledge base
- • Automatic detection and merging of duplicate memories
- • Inter-agent coordination via event notifications
Not For
- • High-security environments without careful secret and access management
- • Organizations that require a fully specified REST/OpenAPI contract and SDKs
- • Workloads needing strong API rate-limit guarantees and documented SLAs
- • Systems that cannot use external LLM/embedding providers (if LLM dedup/chat is enabled)
Interface
Authentication
Authentication appears to be primarily configuration via environment variables for upstream services (D1/S3/OpenAI). No user-facing auth/session management is described for the MCP server itself.
Pricing
Package is MIT-licensed and installable via pip from the repository; pricing for underlying services depends on your configured providers.
Agent Metadata
Known Gotchas
- ⚠ Embedding model changes require rebuilding embeddings and cross-references (manual operational step).
- ⚠ LLM deduplication/chat may fail if the configured OpenAI-compatible endpoint/model does not support required tool/function calling behavior.
- ⚠ Cloud DB modes rely on correct environment variables/credentials; misconfiguration can prevent storage access.
Alternatives
Full Evaluation Report
Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for memora.
AI-powered analysis · PDF + markdown · Delivered within 30 minutes
Package Brief
Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.
Delivered within 10 minutes
Score Monitoring
Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.
Continuous monitoring
Scores are editorial opinions as of 2026-03-30.