Agentic memory framework for LLMs and AI Agents

In today’s newsletter:

• MemU: Agentic memory framework for LLMs and AI Agents

• Firecrawl Skill: Real-time web context for AI Agents

• Agent Squad: Lightweight framework for orchestrating multiple AI agents

Reading time: 3 minutes.

Agentic memory framework for LLMs and AI Agents

MemU is an open-source agent memory framework that lets LLMs store, organize, and reason over long-term memory using a file-system based design.

Instead of stuffing context or relying only on vector search, MemU lets agents read and reason over memory files directly.

Memory is not an index. It’s something the model can understand.

MemU ingests multimodal inputs, extracts structured textual memory items, and autonomously organizes them into thematic Markdown files.

Here’s how memory is structured:

Raw resources → memory items → memory category files

Documents, conversations, images, and audio are preserved in their original form, without deletion or modification. Facts are then extracted and organized into human-readable memory category files.

Key features:

• Dual-mode retrieval, including LLM-based (non-embedding) search for higher accuracy

• File-system based memory where each category is a Markdown file

• Hierarchical memory layers that preserve traceability

• Native multimodal memory for text, images, audio, and video

• Lightweight and developer-friendly, no heavy graph constraints

• Fully configurable prompts for high extensibility

It’s 100% open source.

​MemU Github Repo​

Real-time web context for AI Agents

Firecrawl adds a CLI and agent skill for pulling web content directly into local files, designed for use with coding agents like Claude Code, Codex, and OpenCode.

Web context is still a weak point for many agents. HTML is noisy, scrapers break, and token budgets get burned on content that doesn’t matter.

Firecrawl addresses this by fetching and normalizing web content locally, returning clean Markdown or HTML that agents can consume efficiently without repeated requests.

It supports:

• Scraping main content only, removing navigation and ads

• Web search with result filtering by source, time, or location

• URL discovery via sitemap parsing and crawl mapping

• Full-site crawling with depth limits and rate control

​Firecrawl Skill →​

Agent Squad

Agent Squad is an open-source framework from AWS Labs that manages multiple AI agents and routes conversations intelligently.

Most multi-agent systems struggle with routing queries to the right agent and maintaining context across agent switches.

Agent Squad solves both.

It uses intelligent intent classification to route queries based on context and preserves conversation history across all agents.

Fully implemented in both Python and TypeScript.

Here’s how it works:

• Define specialized agents for different domains (travel, weather, support, health)

• Agent Squad classifies each input and routes it to the appropriate agent while maintaining conversation history across switches

• Supports both streaming and non-streaming responses

Agent Squad Github Repo →​

That’s all for today. Thank you for reading today’s edition. See you in the next issue with more AI Engineering insights.

PS: We curate this AI Engineering content for free, and your support means everything. If you find value in what you read, consider sharing it with a friend or two.

Your feedback is valuable: If there’s a topic you’re stuck on or curious about, reply to this email. We’re building this for you, and your feedback helps shape what we send.

WORK WITH US

Looking to promote your company, product, or service to 160K+ AI developers? Get in touch today by replying to this email.