Memory
Give your agent durable memory so decisions, preferences, and facts survive across chat sessions.
Overview
Velixar memory is a Build-phase MCP server that provides persistent store-search-recall memory for AI assistants across sessions.
What is this MCP server?
- Store, search, and recall information across separate agent sessions
- npm package velixar-mcp-server at version 0.2.4 with stdio transport
- Requires VELIXAR_API_KEY (free key available at velixarai.com per env description)
- Purpose-built persistent memory layer, not ephemeral chat history
- GitHub source at VelixarAi/velixar-mcp-server
- npm velixar-mcp-server version 0.2.4
- required secret VELIXAR_API_KEY
Community signal: 1 GitHub stars.
What problem does it solve?
Agents forget prior decisions every new chat, forcing you to re-explain stack choices, prefs, and open tasks.
Who is it for?
Solo builders running multi-session agent work who want a managed memory API without building their own vector store.
Skip if: Projects that keep all state in git and docs, or teams blocked on sending conversation facts to a third-party memory API.
What do I get? / Deliverables
After setup, the agent can persist, search, and reload memories through Velixar instead of relying only on the current context window.
- Cross-session memory store and search via MCP tools
- Recall of saved facts into new agent conversations
- stdio MCP connection at package version 0.2.4
Recommended MCP Servers
Journey fit
Build is the primary shelf because persistent agent memory is installed while you wire tooling for long-running products. Agent-tooling covers external memory services that extend what the model can recall beyond the context window.
How it compares
Hosted persistent memory MCP, not an in-repo notes skill or local SQLite-only recall hack.
Common Questions / FAQ
Who is Velixar memory for?
It is for developers who want cross-session recall for coding agents via Velixar’s memory API and MCP stdio server.
When should I use Velixar memory?
Use it when the same agent must remember user context, decisions, or facts across days of Build and Operate work.
How do I add Velixar memory to my agent?
Install velixar-mcp-server from npm, configure stdio MCP, and set the required VELIXAR_API_KEY environment variable.