The Recallium blog.
Insights on AI memory, developer tools, and building intelligent systems.
All posts
Guides and field notes from the team, alongside community write-ups from people running Recallium with their own agents. Filter by topic.
How to Stop Coding Agents Re-Reading the Same Files
Every new session, your coding agent greps and re-reads the same files. Why it happens, and six fixes that work in Claude Code, Codex and Cursor.
Serving Memory vs. Engineering Memory
Mem0, Supermemory, Zep and Letta remember your users. Your engineering team keeps losing why the code is the way it is. Two different products, one word.
Can AI Coding Agents Replace Developers?
No. The bottleneck is context, not capability. What developers carry that agents don't, why Big Tech is rehiring, and what has to change before it flips.
How I Coordinate 19 AI Agents Across 4 Model Families With OpenCode and Recallium
“They don’t talk to each other. They share a brain instead.” Running 19 specialized AI agents in production across Claude, GPT, and Gemini - OpenCode handles orchestration while Recallium provides persistent, cross-project searchable memory.
Stop Re-explaining Context to Your AI Coding Agents
Every session you repeat the same conventions and decisions to your coding agent. Shared memory ends the loop for you and your team.
MiniMe-MCP: The AI Memory Layer You've Been Missing
Skywork reviewed Recallium when it was still called MiniMe MCP: how persistent, shared memory ends context reconstruction, cross-project blindness and architecture amnesia.
The Memory Layer Your Coding Agents Have Been Missing
Each coding agent now remembers a little, in its own silo. Why a shared memory layer beats a bigger model, and what it looks like for a team.