Getting Started

AI agents that know your project. One shared brain across Claude Code, Codex, Gemini, Aider, OpenCode, Qoder, and Kimi Code.

0dai generates native configs for supported agent CLIs from a single ai/ directory and keeps their shared context, decisions, and outcomes in sync.

Install

curl -fsSL https://0dai.dev/install.sh | sh
# needs Node.js 22.12+, no sudo

The installer downloads the published CLI bundle, checks its SHA-256 checksum, and puts the launcher in ~/.local/bin/0dai. If that directory is not on your PATH, it prints the line to add.

Start locally — no account required

cd your-project
0dai init --local
0dai run "add a test" --now

0dai init --local creates project context without a 0dai account or activation. 0dai run --now contacts the 0dai API to split your goal into tasks, then runs your installed agent CLI locally. That agent uses its own authentication.

Connect an account when you need cloud features

0dai auth login
0dai activate free
0dai init

Use this path for account-bound setup and cloud sync. Local initialization remains available independently.

Project context

This auto-detects your stack (Next.js, FastAPI, Flutter, Go, etc.) and generates:

ai/ — manifests, personas, skills, playbooks, delegation policy
.claude/ — settings, agents, hooks, rules
.codex/ — config, agents
.gemini/ — settings, agents
.aider/ — config, agents
AGENTS.md, .mcp.json

Verify

0dai doctor    # check health
0dai detect    # show detected stack
0dai status    # maturity, swarm, session

Keep in sync

0dai sync

Run after editing files in ai/ to regenerate all native configs.

What's next

Application clients and orchestration — client direction, capabilities, and delivery status
Session Roaming — transfer context between agents
CLI Reference — all commands
Troubleshooting — fixes for the first errors you hit
MCP Tools — full reference and plan access
Delegation Policy — haiku/sonnet/opus routing
Loop Presets — 8 scheduled autonomous scans

Compare

Why 0dai — shared repo memory and agent config sync
Session Roaming — transfer context between agents
Multi-agent workflow — coordinate parallel agent work
vs Cursor — repo-owned context for multi-CLI teams
vs Aider — pair-programming CLI plus shared memory
vs Claude Code subagents — feature matrix and import adapter

FAQ

Is this just another CLAUDE.md generator?: No. Native configs are the entry point. The value is the living memory, routing, outcomes, and gates around the agents.
Does source code leave the machine?: The free local workflow keeps source local. Product metadata, manifests, and promoted patterns are the cloud boundary.
Why show pricing on the landing page?: The concept is explicit: free stays free, Pro is for the network, Team is for shared operational memory.