Local-first context for AI-assisted work

Where AI sessions become shared work.

Turn private coding conversations into project context your team can act on: tasks, decisions, progress, evidence, and a clear next step.

Remember Decisions, files, and next actions

Coordinate Projects, tickets, and handoffs

Share safely Work state, not conversations

See the workflow

Local-first Open source macOS + Windows 4 AI CLIs

Project workspace

Context Workspace project board with work organized by status

Shared workspace Project state travels through Git. AI conversations stay local.

From chat to project

Your AI chat is not the project.

The conversation creates momentum. Context Workspace preserves it as work your team can understand and continue.

01 Start here

AI session

Ideas, decisions, and code.

02 Turn it into

Clear work

Tasks, context, and next steps.

03 Manage it as

A real project

Board, progress, and effort.

Know what happens next

One board.
Every next step.

See what to do now, what is blocked, and what shipped across every AI session.

Prioritized tasks
Project progress
Recommended next action

Context Workspace Kanban board organizing session work into a managed project

Measure the work

See what the result really took.

Four answers. No guesswork.

Time

How long?

Effort across every session.

Progress

What moved?

Open, active, blocked, and done.

AI usage

What did AI use?

Models, context, tokens, and credits.

Evidence

What changed?

Files, decisions, and session history.

Built-in continuity

Stop anywhere.
Resume ready.

Your context, next action, and project state are ready when you return.

Context Workspace

Work from the CLI.

An AI-assisted agile partner that prepares the evidence and work while you own priorities, acceptance, and decisions.

Project management

Plan and deliver the work

/cw:project /cw:archive /cw:auto-wrap /cw:refine /cw:plan /cw:work /cw:sync /cw:review /cw:retro

Session continuity

Stop and resume with context

/cw:wrap /cw:handoff /cw:reopen

Existing `/sham:*` commands remain available during the compatibility period. Open source on GitHub.

Install with one prompt

Let your AI set it up.

GitHub Copilot · Claude Code · OpenAI Codex · Google Gemini

AI installation prompt

Install Context Workspace from https://github.com/OAbouHajar/smart-ai-human-manager on this machine.

Detect the operating system first. On macOS, verify git, Node.js 22.13+, and at least one supported AI CLI, then run `./scripts/install.sh --no-open`. On Windows, also verify PowerShell 7 and run `pwsh -File .\scripts\install.ps1 -NoOpen`. Stop on unsupported systems.

Clone the latest main branch into a temporary directory, read the README and matching installer, preserve existing Context Workspace data and unrelated AI CLI settings, and configure every detected provider. Verify `http://127.0.0.1:43120/api/health` returns `ok: true`, open the dashboard, and report the installed version, configured providers, and any remaining restart or trust action. Proceed autonomously and only ask before administrator-required or destructive actions.

Local-first

Your work stays yours.

Runs on localhost. Stores data locally. No hosted dashboard.