Best for
- Agent applications requiring cross-session memory
- Users who prioritize data privacy and prefer local memory storage
L
Lumera Agent Memory
Community
Durable agent memory with Cascade object storage and local FTS index
- Category
- Other #226 of 230
- Stars
- ★ 2.8k Very popular
- Transport
- stdio (local process)
- Runtime
- Python
- Credentials
- No credential needed
- License
- MIT
- Last commit
24FMRS · D
This server provides persistent memory for agents, featuring client-side encryption and local full-text indexing, suitable for privacy-conscious and cross-session memory scenarios.
Strongest · Maintenance 7/20 Weakest · Documentation 3/20
Why each score
- Reliability 5/20
- The manifest declares a PyPI package lumera-agent-memory 1.0.0 with stdio transport, and the README gives only a one-line description. No server source, tool list, tests, or CI evidence is supplied, so the MCP init handshake and declared-tool behavior cannot be verified. Missing execution evidence forces a low score under the static calibration rules.
- Security and permissions 5/20
- No evidence covers credential handling, least privilege, external network behavior, or data boundaries. The README's claim of client-side encryption and Cascade object storage is unverifiable from source. No red-line issue such as malware or exfiltration is visible, but there are also no auditable permission controls, confirmation flows, or data-flow disclosures, so the score remains low.
- Maintenance 7/20
- Repo-level facts include an MIT license, an unarchived repository, and ongoing release/validation activity, but these are repository-wide signals. For the lumera-agent-memory server itself there is no commit history, release cadence, dependency update evidence, or security-response channel. With only a 1.0.0 version and no server-specific governance evidence, full credit cannot be given.
- Documentation 3/20
- No server-specific documentation exists beyond a one-line README summary mentioning durable memory, Cascade object storage, and a local full-text index. Install, auth, tool parameters, limits, cost, and troubleshooting are all absent, so this dimension is far below the 'usable but with gaps' anchor.
- Setup experience 4/20
- The manifest provides a package identifier and stdio transport, giving a partial install path, but it omits Python version requirements, environment variables, client configuration examples, platform compatibility, and dependencies. The setup path is therefore unclear and unverified, warranting a conservative score.
Static review · not runListed 2026-08-07
Read the FMRS scoring method →Fit and risk
What it can accessWrites / deletes local filesUses the networkConnects to a database
Not for
- Applications needing complex relational queries
- Scenarios requiring large-scale distributed storage
Required permissions
- Access local filesystem to store and retrieve memory data
- May access Cascade object storage service (specific permissions depend on deployment config)
Risks and side effects
- Memory data may contain sensitive information; client-side encryption helps mitigate leaks
- Local index dependency; storage capacity may impact performance
Setup
Before you start
Runtime：Python
Install the Python package 'lumera-agent-memory' version 1.0.0 via pip. After installation, configure the MCP server in Claude Code to point to the package executable.
Check that it works
After installing (pip package lumera-agent-memory via stdio transport, or through the marketplace), confirm the lumera-agent-memory server appears as connected in your MCP client, then invoke a memory store/recall action and confirm it persists.
Troubleshooting
- Check that 'lumera-agent-memory' is correctly installed in the Python environment
- Ensure the MCP server configuration command path points to the correct executable
- Verify object storage connection settings (if using remote storage)
Things to try
Once connected, you can ask your AI assistant things like:
- Remember the key decisions from this session so we can pick up next time
- What do you remember about this project? Summarize your stored memory
- Store these architecture notes in durable memory with a full-text index
Use cases
Provide persistent memory for Claude Code across sessions
Save and retrieve agent context
Search historical memory using local full-text index
Supported clients
Claude
Listed from the project's documentation, not tested by this site.
Overview
Lumera Agent Memory is an MCP server providing persistent memory for AI agents. It uses Cascade object storage and a local full-text search (FTS) index, supporting cross-session context saving and retrieval with client-side encryption. The server communicates via stdio and is distributed as a Python package.
Similar servers
Hevy MCP
77 · B
Manage your Hevy workout data from AI assistants via MCP.
Firecrawl MCP Server
75 · B
Firecrawl's official MCP server for web search, scraping, and structured extraction for AI agents
BioMCP
75 · B
One binary. One grammar. Evidence from the biomedical sources you already trust.
Compartment
75 · B
Encrypted, fully offline long-term memory for AI agents
More servers from this repository
Source revision ebea1067983c Data synced 2026-10-12 Read the FMRS scoring method