whereismyai
Adaptive Agent Memory Research
Adaptive Agent Memory Research is an open-source research engine providing local-first, model-agnostic long-term memory for LLM agents.
Adaptive Agent Memory Research is a local-first, model-agnostic long-term memory system for LLM agents that decides what to retain based on the measured future usefulness of past experience. It features an append-only log, a controller, and retrieval tools that integrate with MCP clients and the command line. The system operates entirely offline without requiring an API key, vector database, daemon, or cloud account. It is designed for developers building LLM agents who need persistent project memory and reliable recall over long histories. What sets it apart is its explicit focus on honest measurement, publishing both positive results and negative findings or retractions from its internal experiments.
Adaptive Agent Memory Research at a glance
Key features
- Tracks project memory via an append-only events log
- Provides eight built-in tools for MCP clients
- Runs fully offline with zero external dependencies
- Executes commands from the command-line interface
- Maintains an episodic log and a semantic store
- Audits repository claims and detects drifted artifacts
- Measures future utility instead of semantic similarity
Use it for
- Give Claude Code persistent project memory
- Audit repository claims for outdated findings
- Run experiments on agent memory systems locally
- Inspect kept memory and assemble context bundles
Alternatives to Adaptive Agent Memory Research
- Tigerless Agent Memory — Maintains agent history in portable, greppable Markdown files without API keys.
- Memori — Reduces LLM token usage and spend by over 97% with targeted recall.
- Agentmemoryrepo — Enables agents to navigate memories like a wiki and share knowledge across sessions.
- Supermemory — Integrates long-term memory into AI coding agents with a single prompt.
- ReMe — It enables local-first, file-native long-term memory sharing across multiple AI agents.
- MemOS — It delivers ultra-persistent memory, hybrid-retrieval, and token savings.
- LongMemory — Provides immutable, temporal, and governed recall instead of simple vector retrieval.
- SimpleMem — Provides semantic lossless compression to reduce token consumption and maintain long-term context.
- memU — It gives every connected agent a shared and inspectable memory layer.
- Cognee — Persistent memory for AI agents using graph, vector, and relational retrieval.
- Mem0 — It provides drop-in persistent memory for AI agents without changing pipelines.
- EverOS — It enables agents to learn from experience and improves over time without manual coding.
All Adaptive Agent Memory Research alternatives