Adaptive Recall
Give Claude Code or Cursor persistent, scored memory and a knowledge graph so agents recall prior decisions across sessions.
Overview
Adaptive Recall is a MCP server for the Build phase that provides adaptive agent memory with cognitive scoring, a knowledge graph, and self-improving retrieval over HTTP.
What is this MCP server?
- Adaptive memory with cognitive scoring to prioritize what the agent should retain
- Knowledge graph linking entities and context across conversations
- Self-improving ML layer that refines recall over time
- Streamable HTTP remote at s1.adaptiverecall.com/mcp (npm package adaptive-recall v1.0.1)
- Requires ADAPTIVE_RECALL_API_KEY (free tier at adaptiverecall.com)
- Server version 1.0.1
- Transport: streamable-http at s1.adaptiverecall.com/mcp
- 1 required secret env var: ADAPTIVE_RECALL_API_KEY
Community signal: 3 GitHub stars.
What problem does it solve?
Coding agents forget prior architecture choices and user preferences every session, forcing you to re-teach context instead of shipping.
Who is it for?
Solo builders running long Claude Code or Cursor projects who need cross-session memory without building a custom vector stack first.
Skip if: Teams that need fully on-prem, air-gapped memory with no third-party API or hosted endpoint.
What do I get? / Deliverables
After you connect Adaptive Recall with an API key, your agent can persist and rank memories so multi-session builds stay aligned with earlier decisions.
- Durable agent-accessible memory store with scored recall
- Knowledge-graph-linked context usable across coding sessions
- Configured remote MCP connection to Adaptive Recall
Recommended MCP Servers
Journey fit
Memory MCPs sit on the agent toolchain shelf because they are installed to extend what coding agents remember while you build and iterate products. agent-tooling is the canonical home for third-party memory layers that plug into MCP clients during development workflows.
How it compares
MCP memory service, not an in-repo SKILL.md workflow or local SQLite notes plugin.
Common Questions / FAQ
Who is Adaptive Recall for?
Indie developers and small teams using MCP-capable agents who want durable, scored memory without building their own RAG pipeline.
When should I use Adaptive Recall?
Use it when a project spans many agent sessions and you keep losing stack decisions, naming conventions, or open TODO context.
How do I add Adaptive Recall to my agent?
Register the streamable-http MCP URL https://s1.adaptiverecall.com/mcp, set ADAPTIVE_RECALL_API_KEY in your client config, and install the npm identifier adaptive-recall per your agent's MCP docs.