Recall™ — Agent OS

Durable memory, audit, and continuity for AI coding agents.

---

The Problem

Modern AI coding agents are stateless goldfish. Every conversation starts cold. Memory is whatever the host platform decides to keep. Multi-day campaigns degrade into "we already discussed this" loops until the context window compacts your history into oblivion.

The Solution

Recall treats the agent like a contractor who clocks in every day. They don't remember yesterday from their own head — they read the project log. The log lives on disk, in a vector brain, and in immutable cloud storage.

Hot tier   → /memories/         (auto-loaded every turn)
Warm tier  → /memories/session/  (current conversation)
Cold tier  → vector brain        (22 MCP tools, semantic search)
Archive    → Azure Blob WORM     (<$1/mo, 7-year retention)

Six Pillars

#	Pillar	What it does
1	Memory	Tiered file + vector memory that survives across sessions
2	Local Compute	Open-weights models on consumer GPUs for summarization
3	Vault	Immutable cloud backup (WORM) — nothing can be deleted
4	Guardrails	Safety rules that prevent destructive actions
5	Brain Backup	Vector DB with 22 MCP tools for semantic recall
6	Continuity	Cold-start protocol + revival docs for session handoff

Works With

GitHub Copilot · Claude Code · Cursor · Cline · Aider · Continue.dev · LangChain · AutoGen · any MCP-compatible client

Quick Start

pip install ai-recallworks
recall serve

Then add to your MCP client config:

{
  "mcpServers": {
    "recall": {
      "command": "recall",
      "args": ["serve"]
    }
  }
}

By The Numbers

Metric	Value
Cloud storage cost	< $1/mo
WORM retention	7 years
Cold-start brief	~3 KB
MCP tools	22
Independent fail-safes	3

Links

• Website: recall.works
• PyPI: ai-recallworks
• MCP Registry: io.github.RecallWorks/recall
• Whitepaper: Read the full whitepaper →

License

MIT — free to read, free to use, free to adapt.

Recall™ trademark application pending.

---

Built by Steve Paltridge · Mortgagetech