folowise-cli

One brain for all your AI coding tools. Manage agents, skills, rules, and hooks from a dashboard — materialize them into Claude Code, Codex, or Gemini with a single command.

What it does

Folo lets you:

• Manage one brain (agents, skills, rules, hooks, context packs) from a web dashboard
• Run any AI coding CLI — Claude Code, Codex, Google Gemini — with that brain materialized as a system prompt
• Share with your team — multi-tenant orgs, roles, brain versioning, audit log
• Install from a marketplace — pre-built packs (TypeScript, Python, Security, etc.)

Install

Quick install (recommended)

curl -fsSL https://folo-api.folowise.com/install.sh | sh

via npm

npm install -g folowise-cli

Requirements

• Node.js 20 or higher

Quick start

# 1. Log in — opens browser for device code approval
folo login

# 2. Pull your brain from the platform
folo sync

# 3. Run Claude Code with your brain injected
folo run -b claude

# Or start an interactive chat
folo chat

Commands

Authentication & setup

| Command | Description | |---------|-------------| | folo login [token] | Device-code OAuth flow, or pass a token directly | | folo logout | Clear auth and cached brain | | folo init [token] | One-shot setup: login + sync + doctor | | folo doctor | Check config, auth, backend binaries | | folo status | Show current auth, org, brain info | | folo config | Show/set platform URL and default backend |

Brain management

| Command | Description | |---------|-------------| | folo sync | Pull brain snapshot from platform (ETag-aware) | | folo export -o file.json | Export your brain as a FoloPackage | | folo import file.json | Import a FoloPackage into your org |

Running AI tools

| Command | Description | |---------|-------------| | folo run -b claude | Materialize brain, spawn Claude Code with it | | folo run -b codex | Same, for OpenAI Codex CLI | | folo run -b google | Same, for Google Gemini CLI | | folo run -b claude --use-api | Force use of API key over local CLI auth | | folo chat | Interactive chat with session tracking | | folo chat --model opus --budget 5.00 | Chat with specific model + cost cap |

Marketplace

| Command | Description | |---------|-------------| | folo marketplace browse | List available content packs | | folo marketplace browse --category security | Filter by category | | folo marketplace install <slug> | Install a pack into your org brain | | folo marketplace installed | List installed packs |

Integrations

| Command | Description | |---------|-------------| | folo integrations | List connected AI backends | | folo integrations --check | Re-detect locally installed CLIs |

Configuration

Config file lives at ~/.folocli/config.json:

{
  "platformUrl": "https://folo-api.folowise.com",
  "token": "...",
  "defaultBackend": "claude",
  "lastBrainEtag": "\"..\"",
}

Override the platform URL (useful for self-hosted deployments or staging):

folo config --set-platform https://stg-folo-api.folowise.com
# or via env var:
FOLO_PLATFORM_URL=https://my-folo.example.com folo sync

Set a default backend so you don't need -b every time:

folo config --set-default-backend claude

File locations

| Path | Purpose | |------|---------| | ~/.folocli/config.json | Your platform URL, token, defaults | | ~/.folocli/brain.json | Last-synced brain snapshot | | ~/.folocli/sessions/<ts>/ | Materialized CLAUDE.md / CODEX.md / GEMINI.md per run |

How it works

Dashboard (manage brain)  →  API (store + version)  →  CLI (materialize + run)
         ↓                            ↓                          ↓
   Marketplace              PostgreSQL + WebSocket     Claude / Codex / Gemini

The CLI:

1. Pulls your brain (agents, skills, rules, hooks, context) from the API
2. Materializes it into the format the target backend expects
   • Claude Code → CLAUDE.md with --append-system-prompt
   • Codex → CODEX_INSTRUCTIONS.md + CODEX_SYSTEM_PROMPT env
   • Gemini → GEMINI.md + GEMINI_SYSTEM_INSTRUCTION env
3. Spawns the backend CLI with the brain injected
4. Reports session start/end to the platform for audit + cost tracking

Authentication options

Folo supports two auth modes per backend:

1. CLI mode — use your locally-installed CLI's own auth (e.g., claude login from an Anthropic Pro subscription). folo doctor auto-registers it.
2. API key mode — paste your provider API key into the dashboard (Anthropic Console / OpenAI Platform / Google AI Studio). Folo stores it encrypted (AES-256-GCM). Use with --use-api flag or automatically when no CLI auth is available. Works on VPS / remote servers where you can't run claude login locally.

Troubleshooting

folo login says "cannot reach platform" Check your network and that the platform URL is right: folo config

folo run -b claude fails immediately Run folo doctor to check the claude binary is on your PATH.

"Invalid or expired token" Run folo login again.

Installed CLI on a VPS and can't authenticate locally Use API-key mode: add your provider key in the dashboard under Integrations → API tab.

Links

• Dashboard: https://folo-app.folowise.com
• Install page: https://folo-app.folowise.com/cli
• API: https://folo-api.folowise.com
• Source: https://github.com/Hsnaimeh/folo

License

MIT — see LICENSE.