ContxOS Guides

How ContxOS gives AI agents a memory that lasts: what each feature does, how to set it up, and what to ask your agent.

Getting started

Connect Claude to ContxOS: Add Persistent Memory as a Remote MCP Server

Add ContxOS to Claude.ai, Claude Code or any MCP client in about a minute: paste one URL, sign in with OAuth, and install the ContxOS commands.

Updated September 29, 2026

Memory

Persistent Memory for AI Agents: Typed, Linked Knowledge Your Agent Can Search

ContxOS gives agents long-term memory made of typed, linked entries (decisions, results, plans, bugs) that they search by meaning and drill into for the full story.

Updated September 29, 2026

Memory

Keep AI Agent Memory Current: Supersession, Conflict Flags and Expiry

Stale memory is worse than none. ContxOS links new decisions to the ones they replace, flags contradictions, and demotes expired knowledge so agents act on what is true now.

Updated September 29, 2026

Memory

Memory Consolidation for AI Agents: Turning Session Notes into Patterns and Principles

ContxOS Dream reviews what your agents recorded, distills recurring patterns and principles, clears stale plans, and reports what is due, so memory gets sharper instead of noisier.

Updated September 29, 2026

Sessions

Session Handoffs: Resume AI Agent Work Exactly Where You Left Off

Close a session with a structured handoff of accomplishments, pending work, blockers and decisions, then type "resume" in a new chat and continue without re-explaining.

Updated September 29, 2026

Workflows

Reusable AI Agent Workflows: Capture a Procedure Once, Run It Every Month

Turn a procedure your agent worked out into a workflow in the knowledge base: flows and steps any agent can find by phrase, follow exactly, and adapt to a new client or entity.

Updated September 29, 2026

Workflows

Run Long AI Agent Jobs Across Sessions and Subagents Without Losing Your Place

ContxOS tracks a workflow run outside the chat: position, verified facts and holds live on the run, so work survives context resets and steps can go to subagents.

Updated September 29, 2026

Developers

ContxOS for Claude Code: Project Memory Beyond CLAUDE.md

Keep CLAUDE.md for always-on rules and put everything else in ContxOS: decisions, bugs, handoffs, workflows and architecture charts, searchable across repos and sessions.

Updated September 29, 2026

Developers

Architecture Atlas: A Map of Your Codebase That Agents Keep Current

Give coding agents an architecture map they read before changing a subsystem. Charts are current or retired, and handoffs flag the stale ones your work touches.

Updated September 29, 2026

Teams

Separate Projects, Shared Memory: Organizing AI Agent Knowledge for Clients and Teams

Keep each client or product in its own ContxOS project, split one project into buckets by app or engagement, and share a project with teammates whose agents read the same memory.

Updated September 29, 2026

Industries

AI Agent Memory for Accounting Firms: Close Procedures That Remember

Accounting is where ContxOS was proven: close procedures, client-specific rules and every judgment call kept as memory the agent follows next month. Pairs with ERP Sync for the books.

Updated September 29, 2026